Publication Index:

Duct Tightness Data

Duct tightness data presented have been gleaned from BAIHP Trip Reports with some supplementary data from the preceding program, the Energy Efficient Industrialized Housing Project.

Testing Protocol

All duct systems tested were in newly manufactured homes using industry standard methods as delineated in the Minneapolis Blower Door and Duct Blaster User Guides and augmented by the Florida Home Energy Rating requirements where appropriate.

Factory Visits and Test Results Summary

FSEC-BAIHP data spans 1996-2003 and includes test results from 39 visits to 24 factories of six HUD Code home manufacturers (Table 5). Researchers conducted tests on 101 houses representing 190 floors[3]. The data is a compilation of test results from standard production duct systems, repaired to improved production systems.

For a description of improvements implemented at a single manufactured housing plant, see Appendix A.

Average, maximum, and minimum duct leakage data are presented in Table 6 and Figure 7 with similar data from a study published by the Manufacture Housing Research Alliance for comparison (MHRA, 2003.) Figures 8-11 show all data points for Qn_total and Qn_Out. for taped systems (Fig. 8 and 9) and mastic sealed systems (Fig. 10 and 11).

For mastic sealed systems (n=132), average Qn_total=5.1% (n=124) with 85 systems achieving the Qn_total ≤ 6% target (Fig. 10). Average Qn_Out=2.4% (n=86) with 73 systems reaching the Qn_out ≤ 3% goal (Fig. 11).

For taped systems (n=58), average Qn_total=8.2% (n=56) with 19 systems reaching the Qn_total ≤ 6% target (Fig. 8). Average Qn_Out=5.7% (n=30), more than double the mastic average, with 5 systems reaching the Qn_out ≤ 3% goal (Fig. 9).

The average Qn_Out found in this data for mastic sealed systems was 2.4%. This correlates with the Manufactured Housing Research Alliance's study which found an estimated average Qn_Out of 2.5% in 59 floors tested after duct repairs at 16 factories (MHRA, 2003). MHRA did not report the total leakage (measured) used to estimate leakage to the outside, leakage for taped, or leakage for systems before repair.

Figure 7. BAIHP Duct Data Averages
Top: Averaged CFM25Total and CFM25Out data show that mastic sealed systems were tighter than taped systems in both total leakage and leakage to the outside.
Bottom: Averaged Qntotal and Qnout data show that mastic sealed systems, on average, met both the total leakage and leakage to outside goals whereas the taped systems met neither.

Figure 8. All Qntotal data points for taped systems. Note that average (black line) is well above target (red line). 19 tape sealed systems met the 6% Qntotal goal.

Figure 9. All QnOut data points for Taped systems. Note that 5 taped systems met the 3% Qnout goal.

Mastic Sealed Duct Systems

Of the 190 floors tested, 132 had mastic sealed duct systems. Researchers conducted 124 CFM25_total tests and 86 CFM25_Out tests.

Total duct leakage only was measured in 44 mastic sealed systems. Of those, 17 did not meet the Qn_total ≤ 6% goal (Table 7). Problems centered on dimensional coordination of duct components and misaligned pre-cut register holes in sub-floor assemblies, incomplete mastic application, imprecise cutting, and incomplete joints (eg tabs not bent).

Both total and outside leakage tests were conducted on 80 mastic sealed systems, of which 58 floors met both the Qn_total<=6% and Qn_Out<=3% goals. The remaining 22 floors were divided into three groups (Table 8):

• Floors that met the Qn_Out but not Qn_total (n=14). Qn_total range: 6.1% to 9.7%
• Floors that met the Qn_total but not Qn_Out goal (n=1) Qn_Out=4.1%
• Floors that met neither goal (n=7).

Six of the 7 floors that met neither goal were tested during two initial factory visits. One of the factories did not pursue BAIHP recommendations and the other is working toward achieving the Qn_total ≤ 6%.

Only leakage to the outside was measured in 7 mastic sealed systems. One floor had leakage too low to measure. All six remaining floors failed to meet the Qn_out goal (Table 9.)

Qn_Out compared to Qn_total

The MHRA study estimates Qn_Out(Table 6) using a measured Qn_total multiplied by the ratio of Qn_Out to Qn_total for a completed house from the same factory.

For example, if a completed house for Factory A was found to have Qn_total=7% and Qn_Out=3.5%, then the Qn_Out estimation factor for incomplete houses at Factory A would be 0.5 (7%/3.5%). The value of Qn_Out to Qn_total ratios found by MHRA ranged from 24%-60% (MHRA, 2003).

Field measurements in new site built homes (Cummings, et al, 2002.) and many of MHRA's field measurements in new manufactured homes show Qn_Out is often approximately half of Qn_total, and in the absence of field data, MHRA used 50% as the multiplier to estimate Qn_Out from the measured Qn_total (MHRA, 2003.) As mentioned earlier, the goal of Qn_total<= 6% originates from applying the 50% rule of thumb multiplier to obtain a Qn_Out<=3% goal, which is the BAIHP recommended duct leakage level corresponding to the most stringent duct leakage level in the Manufactured Home Energy Star program.

BAIHP data includes 26 taped systems that researchers tested for both total and outside leakage. The average ratio of outside to total leakage was 56%, roughly agreeing with the rule of thumb. However, in the 80 mastic sealed systems, the average ratio of outside leakage to total leakage was somewhat lower than expected at 36%. There were 13 mastic sealed systems that met the Qn_out ≤ 3% goal without meeting the Qn_total ≤ 6% goal.

This lower than expected ratio is perhaps due to the improved sealing at joints between duct components but not between the house envelope (e.g. subfloor or ceiling) and the air distribution system (e.g. supply boots and return plenums). Leakage where the supply boot joins to the house is part of the total leakage, but tends to be associated with leakage to the interior of the house.

Though the average ratio of outside to total leakage in the mastic sealed systems was slightly lower than expected (36%), the range spanned 0% (leakage to outside too small to register) to 80%. The data strongly supports that achieving a Qn_total of 6% signifies that the Qn_Out will be less than 3%. One exception was documented (Qn_Out=4.1%), proving that using Qn_total as a surrogate test for the Qn_Out goal is not a guarantee.

Quality Control: Tangible Success

An objective quality control strategy is essential to achieving tight duct construction. If air were visible to the naked eye, a visual inspection would reveal leakage sites in any given duct system. In the absence of visible air, managers and line workers will need to learn a way to evaluate their duct construction quantitatively using pressure testing equipment common to building science.

Initially, a standardized duct test on the factory floor provides an objective evaluation of current practice, repairs, and process improvements. Ultimately, pressure testing all duct systems replaces subjective evaluation with a tangible, objective measure of success: total duct leakage, CFM25_total or a ratio of duct leakage to conditioned area, Qn_total. These surrogate measurements are shown by this data and other field studies (Cummings, et al, 2002. MHRA, 2002) to substantially correlate with duct leakage to the outside of completed houses, the factory's ultimate quality goal.

The support of BAIHP as objective, third party experts is often cited by manufacturers as a major benefit. Some manufacturers have already adopted the test procedure into their production process to conduct their own in-house verification of duct system tightness. This leads to a higher quality product as well as accountability of both the factory and field work force.

Continue to Economics of Duct Tightening in Manufactured Housing >>