The following article was published in ASHRAE Journal, March 2005.

© Copyright 2005 American Society of Heating, Refrigerating and

Air-Conditioning Engineers, Inc. It is presented for educational purposes only. This article may not be copied and/or distributed electronically
or in paper form without permission of ASHRAE.

ASHRAE Standard 152

& Duct Leaks in Houses
By Mark Modera, Ph.D., P.E., Member ASHRAE

D
o ducts in houses leak? Tests have shown that in most houses
the answer is yes, and most of those ducts leak enough to merit
sealing. Duct leaks create uncontrolled airflows with consequences house required duct improvement ser-
vices, as well as while the dealer was
that include low pressure zones, increased infiltration that can increase performing duct improvement services
or decrease humidity, nonuniform temperatures, and energy/capacity in houses that needed them.
Dealers performed a house pressure
losses for the HVAC system. test, which measures the change in house
pressure associated with turning on the
This article discusses data collect- the flow through those ducts, and of the HVAC system fan. The test distinguishes
ed from thousands of houses around type of HVAC equipment to which they houses that really need duct sealing,
the country and also discusses ANSI/ are connected. In addition, the standard those that might benefit from sealing,
ASHRAE Standard 152–2004, A Method allows comparison of the efficiency of and houses in which the cost of sealing
of Test for Determining the Design and using hydronic baseboards vs. installing is probably not justifiable.
Seasonal Efficiencies of Residential a duct system in a new house. Supply leaks to the outdoors cause the
Thermal Distribution Systems. house pressure to decrease when the fan
The standard quantifies how much Duct Performance Data is turned on, while return leaks increase
energy and HVAC equipment capacity Over the past decade or so, increasingly house pressure. As a result, the house
duct leaks actually waste. In addition, more groups have measured field perfor-
it calculates the energy and capacity mance of residential HVAC systems.1–7 About the Author
implications of where the ducts are lo- The previously unreported data presented Mark Modera, Ph.D., P.E., is vice president for
strategic operations at Carrier-Aeroseal, Syracuse,
cated (attic vs. crawlspace vs. basement, here was collected by residential HVAC N.Y.; a staff scientist at Lawrence Berkeley National
as well as Miami vs. Minneapolis), of dealers as part of a diagnostic test to de- Laboratory, Berkeley, Calif.; and was chair of Stan-
the insulation level of those ducts, of termine whether a particular customer’s dards Project Committee 152.

28 ASHRAE Journal ashrae.org March 2005

 Duct leaks create uncontrolled airflows with consequences

that include low pressure zones, increased infiltration that

can increase or decrease humidity, nonuniform temperatures,

and energy/capacity losses for the HVAC system.

pressure test detects leakage only to outside—to or from a grille in a house. The results in Figure 2 represent the sum
vented attic or crawlspace. of each house’s register flows divided by the nominal cool-
The situation becomes more complicated when both supply ing capacity of the HVAC equipment. These results can be
and return ducts are leaking. In this case the pressure change compared to a nominal flow of 400 cfm/ton (54 L/s per kW),
is proportional to the difference between the two leakage rates, which is where most residential air conditioners are rated for
not the sum of the leakage rates (e.g., the pressure change would efficiency and capacity.
be zero if they are both leaking the same amount, even if the The significant difference between the actual flows in Figure
leakage flows are very large). 2 and 400 cfm/ton (54 L/s per kW) stems predominantly from
To distinguish tight ducts from equally leaky ducts, the two effects: 1) supply duct leakage, and 2) excessive resistance
house pressure test includes a second step: partially blocking to airflow. The particularly low register flows in the Illinois
the return grilles, increasing the pressure differences across data are most likely due to the larger quantity of leaks in the
the return leaks and decreasing the pressure differences across rectangular sheet-metal duct systems typically found in that
the supply leaks. If the ducts are tight, changing their pressures region (see Figure 4 and associated discussion).
does not affect the house pressure. However, if supply leakage As noted earlier, fan flow, duct leakage and duct insulation
was masking return leakage, the return-blocked test will make all impact the temperature of the air leaving the grilles. Figure
the return leaks more prominent. 3 shows the results of register temperature measurements in
One complication is that the leakage of the shell of the house almost 2,000 houses. The data presented are based upon tem-
determines the magnitude of the pressure change associated perature measurements at all supply and return grilles.
with a given duct leakage flow. For that reason, this test is not The percentage loss in sensible capacity at any supply grille
a precise measurement of duct leakage, but an indicator that is calculated by dividing the temperature difference between
only gives one of three answers, as shown in Figure 1. that grille and the best supply grille, by the temperature dif-
Figure 1 shows results of approximately 1,000 house pressure ference between the best supply grille and the average of the
tests in Sacramento, Calif., and Austin, Texas. Both regions are return grilles (Equation 1).
predominated by attic duct installations of flex-duct systems.
Figure 1 demonstrates two interesting points: 1) between 60% % Thermal Loss = (Tbest-grille – Tgrille)/(Tbest-grille – Treturn-average) × 100
and 85% of the houses in these regions would benefit from duct (1)
sealing, and 2) the distribution of results is remarkably similar This procedure estimates the sensible capacity using the
in the two regions (we also found similar consistency between difference between the average return temperature and the best
dealers in each region). supply temperature, which in most cases is an underestimation
Another important aspect of duct performance is how much due to leakage and conduction losses between the return grille
air actually is delivered to the rooms being conditioned. Fig- and the air handler, and between the HVAC equipment and the
ure 2 shows results of another diagnostic test performed by best supply grille. These results are striking, as they indicate
the HVAC dealers to measure the airflow delivered by each that, on average, between a quarter and a third of the registers

March 2005 ASHRAE Journal 29

 350
Sacramento Avg. (783 Tests) Austin Avg. (199 Tests) C
300
A
B
<15% 250
<15%

Flow (cfm/ton)
200

>25% >25% 150

15% – 25% 15% – 25%
100

50

0
Sacramento Sacramento Sacramento Bakersfield Illinois
2,159 Houses 1,898 Houses 556 Houses 47 Houses 42 Houses

Figure 1 (left): Distributions of house pressure test results, where percentages are combined supply and return duct leakage as a fraction
of HVAC fan flow. Averages are for all participating dealers in the region. Figure 2 (right): Register flow measurement results. A, B, and
C refer to different dealers.

4.5 C 900
4.0 800

Leakage (cfm @ 25 Pa)

3.5
Number of Registers

700
B
3.0 A 600
2.5 500
A
2.0 400 C
B
1.5 300
1.0 200 New
Construction
0.5 100
0.0 0
Sacramento Sacramento Sacramento Bakersfield Illinois Sacramento Sacramento Sacramento Bakersfield Illinois
525 Houses 960 Houses 374 Houses 32 Houses 26 Houses 443 Houses 279 Houses 110 Houses 1,231 Houses 1,065 Houses

Figure 3 (left): Results of register temperature measurements (registers with greater than 25% thermal loss are within a given house). A, B,
and C refer to different dealers. Figure 4 (right): Results of fan-pressurization tests of duct leakage. A, B, and C refer to different dealers.

are losing more than 25% of the sensible capacity provided and measured fan flows to calculate duct leakage as a fraction
by the equipment. Moreover, whether the loss is caused by of fan flow. Also the house would need to be simultaneously
leakage, low flow, or poor insulation, these calculated losses pressurized to the same pressure as the ducts to isolate duct
represent only the thermal conduction loss on the supply side, leakage flows to/from outside the house.
which would apply to the flow being delivered after the supply Figure 4 suggests some interesting trends. First, the three
leakage has been subtracted. dealers in Sacramento all measured approximately the same
The data presented in Figures 1 through 3 was obtained level of leakage, which one would hope to find. This is consis-
during diagnostic services, while Figure 4 presents the tent with the house pressure test results. Second, the leakage
results of fan pressurization tests of total (i.e., supply plus levels in new Bakersfield, Calif., construction are considerably
return) duct leakage performed just prior to sealing in almost lower than the data from existing houses in Sacramento. This
3,000 houses. also is consistent with the assumptions within the California
As compared to the house pressure test, these results are Energy Efficiency Standards about duct leakage in new vs.
more precise and repeatable, although they represent a differ- existing buildings. However, the same dealer installed the duct
ent leakage characteristic, namely the flow through all supply systems being sealed, and, therefore, had an incentive to mini-
and return leaks at a fixed reference pressure. The unit, cfm at mize leakage introduced during the installation process.
25 Pa (0.1 in. w.g.), represents the leakage flow at a reference The other key trend in Figure 4 is that the data from Illinois
pressure differential of 25 Pa (0.1 in. w.g.) (considered repre- shows significantly higher leakage levels, which is consistent
sentative of the average pressures across leaks during normal with other studies of rectangular sheet-metal duct systems
operation) created by the duct pressurization fan. in basements.*
It also represents a combination of SI and I-P units, and This type of duct construction has much more extensive
unfortunately has become the typical unit of reporting. To
*A comparison of the leakage data in Strunk, et al., 1996, with Jump and Modera,
compare these leakage values with house pressure test results, 1994, shows that the floor-area-normalized duct leakage flow at 25 Pa (0.1 in. w.g.) was
they must be combined with measured operating pressures more than three times as much in the basement houses.

30 ASHRAE Journal ashrae.org March 2005

 100% 45%
90% 3%/3% R-4.2
40% 15%/15% R-4.2 Crawl
80% 3%/3% R-38
35%
70%
60% 30%

50% 25%
40%
20%
30%
15%
20%
10%
10%
0% 5%
Heating, Heating, Cooling, Cooling,
Design Seasonal Design Seasonal 0%
15%/15% R-4.2 3%/3% R-4.2 Heating, Heating, Cooling, Cooling,
3%/3% R-8 3%/3% R-38 Design Seasonal Design Seasonal

Figure 5 (left): Application of Standard 152 to an attic flex-duct system in Sacramento shows the attic duct efficiency. Figure 6 (right):
Savings calculated with Standard 152 for changes to an attic flex-duct system in Sacramento.

return duct systems and many more joints as compared to plastic project committee in 1993 that ultimately developed Standard
flex-duct systems; common practice in residential basement 152.
construction does not include sealing those joints. Standard 152 was developed to provide a means for rating
In addition, a small fraction of the Illinois houses had the performance of different thermal distribution systems.
building cavity returns, which generally have higher leakage Its scope was limited to air systems, hydronic systems, and
as compared to sheet-metal systems. One caution relative to electric systems.
all of the data in Figure 4 is that some measurements include Minimum inputs for rating a duct system are:
both supply and return leakage, whereas other measurements • Duct leakage (supply and return);
include only one side of the fan. Based upon conversations with • Duct location (e.g., attic, crawlspace, basement);
the dealers who took the data, • Duct insulation level;
the Illinois data contains many 100% • House location (from a list
more single-sided measure- 90% of cities); and
ments. This suggests that the 80% • HVAC equipment char-
ratio of leakage areas between 70% acteristics (type [e.g. heat
rectangular sheet-metal base- 60% pump], capacity, fan flow).
ment ducts and attic flex-duct 50% The outputs are:
systems is underestimated by 40% • Heating design efficien-
this data. 30% cy;
Something else to keep in 20% • Heating seasonal eff i-
mind when comparing attic 10% ciency;
and basement construction is 0% • Cooling design efficiency;
Heating, Heating, Cooling, Cooling,
that the pressure differentials Design Seasonal Design Seasonal and
seen by the leaks in basement 30%/30% R-1.5 5%/5% R-1.5 5%/5% R-8 • Cooling seasonal eff i-
systems are typically smaller 30%/30% R-1.5 Wall Insulation 30%/30% R-1.5 Ceiling Insulation ciency.
than those in attic systems. Based upon the inputs,
Figure 7: Application of Standard 152 to a basement system in
The net result is that a reason- Chicago shows the basement duct efficiency. Standard 152 first calculates
able estimate of leakage for the fraction of the condition-
attic flex-duct systems is 15% to 20% of fan flow on each side ing produced by the HVAC equipment that is delivered at the
of the fan,3 while rectangular sheet-metal duct systems have supply registers. The standard calculates this fraction, called
more like 25% to 30% leakage on each side of the fan. the delivery effectiveness, using fixed algorithms to calculate
the temperatures in each duct zone using the local climate
ASHRAE Standard 152 conditions.
As duct leakage and temperature loss data began to appear in The local climate conditions are ASHRAE Handbook design
the early 1990s, concerns quickly arose about how to interpret values for the design efficiencies, while the seasonal climate
that data. These concerns led to the formation of a standards conditions are based upon load-weighted seasonal averages of

March 2005 ASHRAE Journal 31

hour-by-hour climate data. The zone-temperature algorithms which accounts for the fact that the extra load associated with
include effects such as the presence of a radiant barrier in an at- duct losses increases the use of electric resistance heating by a
tic, and the location of wall and ceiling insulation in basements. heat pump. At a delivery effectiveness of 0.75, the equipment
For example, Standard 152 temperatures for a well-vented attic interaction factor reduces the distribution efficiency by 14%.
are Tdesign + 22°F at design conditions, and + 13°F for
seasonal efficiency calculations. Example Applications of Standard 152
Standard 152 then calculates the overall distribution effi- Some examples of efficiencies calculated with Standard
ciency, adjusting the delivery effectiveness by the fraction of 152 are illustrated in Figures 5 through 7. Figure 5 shows dis-
energy losses that are recovered into the conditioned space. The tribution efficiency results for Sacramento at different levels
regain factors are based upon the ratio of the thermal conduc- of duct leakage and insulation. It shows that efficiencies for a
tance between the duct zone and the conditioned space, to the duct system with R-4.2 duct insulation and the leakage levels
overall thermal conductance of the duct zone. Typical regain observed in the field for existing houses (15% supply, 15%
values are 10% for a vented attic, 50% for an uninsulated base- return) range from 53% for cooling under design conditions
ment, 75% for a basement with insulated walls, and 30% for a to 75% for the average value over the heating season. Figure 5
basement with an insulated ceiling. also shows that sealing the duct leakage has the largest impact,
Standard 152 also accounts for interactions between HVAC and that the combination of sealing and super-insulation of the
equipment and ductwork, including the effect of ductwork ducts (e.g., burying them in loose-fill insulation) can bring all
losses on equipment operating mode, as well as the effect efficiencies above 90%.
on equipment efficiency of low fan flows. For example, the Figure 6 compares the savings that can be achieved by differ-
equipment interaction factor (which multiplies the delivery ent duct system improvements in Sacramento and demonstrates
effectiveness) for a heat pump is: that moving the ducts from the attic to the crawlspace has little
impact on heating, but produces cooling savings similar to those
Fequip = 0.44 + 0.56 × (Delivery Effectiveness) (2)
achieved by sealing the attic ducts. Figure 6 also illustrates the

Advertisement formerly in this space.

32 ASHRAE Journal ashrae.org March 2005

dramatic impacts of duct improvements on cooling efficiency at
design conditions, which translates to comparable impacts on
overall cooling system capacity at design conditions.
Figure 7 presents duct efficiencies for basement ducts in
Chicago. At typical unsealed leakage levels of 30% supply
and 30% return for an uninsulated metal duct system in an
uninsulated basement, cooling efficiencies still are considerably
higher than for attic ducts in Sacramento, due to the coolness of
the basement.8 Figure 7 also shows the impacts of duct sealing
and insulation, as well as the impacts of basement insulation
and location.

Measurement Procedures in Standard 152

In addition to standardizing calculation procedures, Standard
152 also specifies two procedures for measuring duct leakage
and two procedures for measuring HVAC system fan flow. The
two duct leakage test procedures both include calibrated-fan
pressurization of the duct system. The first procedure pressur-
izes the house with a blower-door fan to the same pressure as the
duct system and measures duct leakage to outside. The second
procedure pressurizes the ducts only, thereby measuring total
duct leakage. The standard only allows the second procedure
to be applied to duct systems having leakage levels of less than Advertisement formerly in this space.
10% of fan flow. For either test, to adjust fixed-pressure duct
leakage flows to the specific systems being tested Standard
152 also specifies procedures for measuring the duct system
operating pressures.
The two procedures within Standard 152 for measuring fan
flow are: 1) a method that involves substituting a calibrated-plate
for the system filter, and 2) a method that involves substituting
a calibrated fan for the return ductwork. Method 2 works by
adjusting the calibrated fan speed to produce the same supply
plenum pressure as found during normal operation.

References
1. Cummings, J.B., R. Dunsmore, and J.J. Tooley, Jr. 1990. “Impacts
of duct leakage on infiltration rates, space conditioning energy use, and
peak electrical in Florida homes.” Proceedings of ACEEE.
2. Jump, D.A. and M.P. Modera. 1994. “Impacts of attic duct retrofits
in Sacramento houses.” Proceedings of ACEEE.
3. Jump, D.A., I.S. Walker and M.P. Modera. 1996. “Field mea-
surements of efficiency and duct retrofit effectiveness in residential
forced-air distribution systems.” Proceedings of ACEEE.
4. Modera, M.P. 1993. “Characterizing the performance of residen-
tial air distribution systems.” Energy and Buildings 20(1):65–75..
5. Modera, M.P. and D.A. Jump. 1995. “Field measurements of
the interactions between heat pumps and duct systems in residential
buildings.” Proceedings of ASME International.
6. Palmiter, L. and P.W. Francisco. 1996. “Measured efficiency
of forced-air distribution systems in 24 homes.” Proceedings of
ACEEE.
7. Strunk, P.R., et al. 1996. “Stock characterization and energy
savings potential in forced air systems in frostbelt homes.” Proceed-
ings of ACEEE.
8. Treidler, E.B. and M.P. Modera. 1996. “Thermal performance
of residential duct systems in basements.” ASHRAE Transactions
102(1): 847–858.
March 2005 ASHRAE Journal 33