Residential HVAC Mythbusting
Two nagging duct questions answered
The idea was most likely born out of watching too many episodes of Mythbusters on the Discovery Channel, but it’s really about the scientific method: When you are presented with a set of supposed facts, put them to the test. In July 2010, Delta-T, Incorporated, a company that formerly specialized in duct sealing (over 5,000 systems sealed) but now focuses on consulting on large-scale utility HVAC programs, ran a series of tests to investigate the veracity of accepted beliefs on the subject of assembling residential ductwork. We carried out two separate experiments, designed to answer two nagging questions. First, what is the best way to connect sheet-metal duct to flex duct? And second, which parts of sheet-metal ductwork is it most beneficial to seal?
Here’s one myth that doesn’t need busting: The sad truth is that some duct installation specifications require specific steps, take extra time to complete, and produce no benefit whatsoever. This waste of time and effort undermines the credibility of the conservation program with installers. So what? They get paid. But interestingly, a job well done, work satisfaction, personal well-being, and a credible program mean more than simply doing "stuff" because the latest utility program roll-out requires it. It’s very easy to assume that since parts of the specifications are superfluous (or just a waste of time), the entire program just might be, too.
QUESTION 1: What is the best way to connect sheet-metal duct to flex duct?
Our method for testing connection strength was to put the assembly in pure tension. This was done by hanging a short section of sheet-metal duct from the ceiling, and then, for each type of connection, attaching a short length of the inner polyester liner of flex duct, followed by another section of sheet metal. Hanging from the lower section of sheet metal was an egg crate. We loaded this crate incrementally with an assortment of dumbbell weights. In this way, we were able to determine a tensile weight to failure for each connection type. It is important to note here that this assembly was in pure tension. That is, the only force acting on the connections was the loaded egg crate pulling straight down. It would seem rare and implausible for an actual duct system to be loaded like this; more likely, the forces would be a combination of torsion and bending. However, as connection failure would ultimately result from tension on the connection itself, the pure tension test provides a reasonable benchmark from which to compare different connection types. Our method was sort of a worst-case scenario for connection failure, which makes the test a reasonable analog for relative connection strength in field conditions.
The different connections we investigated included hand-tightened tension ties, tool-tightened tension ties, tension ties with screws inserted to hold them in place, and a worm drive. With each connection type, we also tested with and without a bead -- the raised nub that runs around the circumference of the sheet-metal duct -- and we also investigated the effects of mastic sealant.
The results of our tests proved unsurprising. Tool-tightened tension ties were the best connection in terms of combined strength and ease of installation (the worm drive was actually a little bit stronger, but it was also more time consuming to install). We did, however, make several interesting notes on the other connections tested. First, it doesn’t matter how strong you think you are. A seven-year-old child with a tension tool will make a stronger tension tie connection any day of the week than a grown man doing it by hand. The difference between hand tight and tool tight was stunning.
Second, inserting screws seemed to make the connection stronger at first, but it became clear that, once punctured, the flex duct gradually unraveled under minimal loading. But the flex duct itself was pretty much indestructible unless we put a hole in it. The flex duct never failed in any of the other tests; the failure mode was always connection slippage -- unless, of course, the flex duct was punctured in some way, in which case the flex duct unraveled every time. The obvious lesson here seems to be that putting screws in flex duct is a really bad idea. Atco Rubber Products, a manufacturer of flex duct, posts the following warning on its web site, in response to the frequently asked question, Is it all right to use screws to help attach flex duct?
ATCO Rubber Products does not recommend screws be used to fasten the polyester core of the air ducts because they weaken the polyester. Polyester is a very strong material as long as there are no holes or tears in it. As soon as a tear or hole is introduced, its strength drops. To maintain our UL approval status, our air ducts must pass a tension test (25 pounds hanging from one end of the duct(, a torsion test (one end rotated 180° or to 25 foot-pounds whichever comes first), and then a leakage test. In all of these tests, both ends must be connected to collars per our installation instructions. Flexible duct connected with screws would not pass these tests.
The other interesting note is that, rather than strengthening the connection, mastic acted as a lubricant and actually reduced connection strength. We ran these tests roughly 24 hours after applying the mastic (giving it a day to set), but because it was wedged firmly between a layer of polyester and the sheet metal, it wasn’t exposed to much oxygen and thus dried extremely slowly. Even at five days of drying time, the mastic acted as a lubricant. One connection (tool-tightened tension tie) was left to dry for two weeks. Its failure point was similar to that of a connection made with the tension tie only. We found that when sealing a sheet to flex connection with mastic, connection failure is very likely; the mastic offers no benefit over connection with a tool-tightened tension tie alone.
So keep it simple. Just use a tool-tightened tension tie; it’s easy to install, and it won’t slip. Put another way, if there are forces large enough to break this connection, then the homeowner probably has a bigger problem than detached ductwork.
Question 2: Which parts of sheet-metal ductwork is it most beneficial to seal?
For round two of the mythbusting, the objective was to determine exactly how valuable it is to seal individual small-leakage parts of sheet-metal ductwork. We all know that it’s crucial to seal dovetailed takeoffs and other such connections, but what about the smaller-leakage parts? For example, what is the potential leakage reduction from sealing just the linear seam? Or just the elbows? Or the butt joints? To test these connections, we needed to construct a duct system in a controlled environment, where we could constantly take measurements, make small adjustments, and take more measurements -- all in a repeatable and meticulously managed setting. To this end, we assembled 75 feet of sheet-metal ductwork with seven 90º elbows and sheet metal-to-flex duct connection points in the lab (my garage, actually). The duct system was constructed with a known number of holes near the end of the duct system that remained unsealed throughout the process.
The parts of the duct system were incrementally sealed, and the system was tested. Blue painter’s tape was used to seal the butt joints, the liner seam, and the flex-to-metal connections. Mastic was used to seal the gores in the elbows. We tested the painter’s tape for airtightness, and it passed with no measurable leakage. Finally, before retesting for leakage, we used chemical smoke to test for leakage while the ducts were under pressurization. Our results are summarized in Table 1.
Table 1. CFM Reduction by Duct System Component at 50 Pa
Butt joints. Even with a better-fitted joint than is typically found in the field, sealing these joints proved worthwhile. The total reduction for the 21 joints was 26 CFM50, or an average reduction of 1.26 CFM50 per joint.
Elbows. For the test, we used 8-inch flexible elbows all set to 90º. We took great care to seal the elbows, and the smoke test proved valuable in finding small gaps in the sealant. Sealing the seven elbows resulted in a reduction of 11 CFM50, or an average reduction of 1.57 CFM50 per elbow. It should be noted that in other tests involving larger elbows, the reduction was greater.
Linear seams. The snap lock joint on sheet-metal ducts is a very tight one, as anyone who as ever tried to dismantle a snapped-together duct knows. The 75 feet of linear seam sealing yielded a total reduction of only 5 CFM50, or an average reduction of .07 CFM50 per linear foot.
Flex-to-sheet metal connections. We used only the inner liner of the flex for this test, since it’s the primary air barrier. The tension ties were tool tightened. Sealing the 14 flex duct-to-metal connections yielded a total reduction of only 3 CFM50.
Conclusions. Within the scope of retrofit duct sealing, the flexible elbow and the butt joints can yield good reductions if done in a diligent manner. Sealing the flex duct to metal connections with mastic yields only a small reduction in leakage. If code requires that this joint be sealed, consider using a UL-listed tape or applying the mastic on the outside of the connection, not under the flex duct. It is important to make sure that each joint is tightened with a tension tool. Linear seams are time-consuming to seal and do not yield much reduction in leakage, especially when the seams are pointed away from the duct sealer.
While our research focused on the small leaks, it is critical to seal all leaks that see high pressure, such as those in the plenum and the furnace base can. It is important for duct-sealing programs to write good specifications backed up with solid training and a good quality control or quality assurance program. Having duct sealers seal ductwork in ways that only add time and frustration to an already physically demanding job will drive up the cost of duct sealing and lower the motivation of the most important person in any duct-sealing program -- the person applying the mastic.
Enter your comments in the box below:
(Please note that all comments are subject to review prior to posting.)
While we will do our best to monitor all comments and blog posts for accuracy and relevancy, Home Energy is not responsible for content posted by our readers or third parties. Home Energy reserves the right to edit or remove comments or blog posts that do not meet our community guidelines.