| Back to Contents Page | Home Energy Index | About Home Energy |
| Home Energy Home Page | Back Issues of Home Energy |
Home Energy Magazine Online September/October 1993
Finding the Energy Culprits
Experts Explain the Techniques
by John Proctor, Michael Blasnik, Bruce Davis, Tom Downey, Mark P. Modera, Gary Nelson, and John J. Tooley Jr.
John Proctor heads Proctor Engineering Group in Corte Madera, California. Michael Blasnik is research director for GRASP in Philadelphia, Pennsylvania. Bruce Davis is senior project manager with the North Carolina Alternative Energy Corporation in Research Triangle Park, North Carolina. Tom Downey is field manager at Proctor Engineering. Mark P. Modera is a staff scientist at Lawrence Berkeley Laboratory in Berkeley, California. Gary Nelson is vice president of the Energy Conservatory in Minneapolis, Minnesota. John J. Tooley Jr. is president of Natural Florida Retrofit Incorporated in Montverde, Florida.
This article is an effort to bring together the ideas of several innovators who have invented methods of diagnosing duct leakage. It focuses primarily on production technology--diagnostic tools that can be used in programs designed to seal tens to thousands of systems per year. Other, more time-consuming measurements exist for research purposes. These methods, along with other factors, are used to predict energy losses due to duct leakage.
Working on duct systems will often change the pressure distribution in a home, sometimes dramatically. These changes can effect combustion appliance drafting, radon migration, moisture, ventilation, and indoor air quality. The diagnostic tools we describe here should be used only by individuals who have a working knowledge of these safety issues and who take precautions to deal with them.
Each of the diagnostic methods can be viewed as a new tool for our toolbox. Some of the tools are for special cases, while others will become the tool we reach for most often. These diagnostic methods can be classified as quantitative or qualitative.
Quantitative methods provide measurements of duct leakage in cubic feet per minute (cfm). They estimate neither the actual leakage from the ducts when the system is operating, nor the leakage across the leakage sites when all sites are at the same pressure. The purpose of a production quantitative measurement is to obtain quality work that will reliably impact energy use.
A good quantitative diagnostic tool has the following features:
With the wide variety of diagnostic tools available, investigators have noted substantial differences in the measured duct leakage between test methods. Variations between test methods are the result of differing pressures across the leakage sites (see Table 1) and different tests measuring different leakage locations (see Figure 1).
Repeatable results--so inspections will produce nearly identical readings.
Accurate information--so technicians can immediately distinguish work resulting in high energy savings from work that has little or no effect.
Quick--so technician time devoted to proper installation or repair of the distribution system is maximized.
The more closely the test conditions match the normal operating conditions of the ducts, the more accurate the test method. Ideally both flow and pressure should be duplicated in the test procedure.
If duct leaks are evenly distributed throughout the system, the pressures in the system are distributed in a manner similar to Figure 2. Leakage location and location of the filter substantially influence the actual pressure distribution. The highest pressures occur at the supply and return plenums. The return system is under negative pressure while the supply system is positive. Reference pressure for the test must be specified as well.
Since early testing was an expansion of blower door testing, the first test pressures were 50 pascals (Pa). There is now a trend toward more accurate pressure measurement (multiple duct locations with a digital manometer) and lower test pressures such as 25 Pa (or .1 in. of water column).
The actual measured operating pressures of the system (when the air handler fan is running) are helpful in interpreting the results of the quantitative duct leakage tests.
The three primary quantitative methods used to measure flow through leaks (at a test pressure) are: blower door subtraction, blower door and flow hood, and Duct Blaster (see Table 1). Other quantitative methods used primarily for research include: tracer gas, tracer temperature, and combined STEM/FAST testing. Tracer gas measures leakage and both tracer temperature and the STEM/FAST combination measure the energy effect of duct leakage and duct conduction.
Blower Door Subtraction Method
The blower door subtraction method estimates flow through duct leaks to outside with the house at 50 Pa. (In all circumstances here, outside means outside the thermal envelope unless noted differently.) Either pressurization or depressurization tests are roughly equivalent (see Two Favorite Test Methods, By the Book, p.32). These will both be referred to as pressurization. This method uses two blower door flow readings to determine the amount of duct leakage. The house is pressurized with a blower door to obtain the total leakage of the structure including the duct leaks. All duct openings are then covered and another blower door reading is taken to obtain just envelope leakage. Both tests are done with the house- to-outside pressure differential of 50 Pa. The total leakage of the second test is subtracted from the total leakage of the first test, yielding the duct system leakage.
Two significant errors are introduced using this method. First, the blower door is measuring relatively large flows (whole house leakage with and without ducts at 50 Pa). Small percentage errors in these readings become large percentage errors when applied to the duct leakage (typically 10%-20% of total house leakage). Second, the method assumes that all of the leakage from the ducts to the outside is eliminated when the registers are sealed. If there is any leakage at the registers, or any other leakage from the house to the duct system, this assumption is incorrect.
The first flaw is critical. An error of 5% in only one of the blower door readings (due to operator error or wind effects) becomes a 50% error in duct leakage for a system with 10% of the house leakage in the ducts. For example, if the initial test shows a leakage of 3,000 cfm50 (which is 150 cfm low), and the second test shows 2,700 cfm50 (with no error), the estimated duct leakage is 300 cfm. The true difference however, is 450 cfm (150% of the estimate).
The second flaw can be overcome. Using a method developed by Michael Blasnik, the error due to leakage from the house to the ducts can be estimated (more on this later).
Flow Hood Method
The flow hood method estimates flow through duct leaks to the outside. During the test, all registers except the largest, least restricted location are blocked. The house is pressurized (or depressurized) to 50 Pa relative to outside and the flow hood is used to measure the amount of air flowing through the open grille. Any flow through the flow hood into the grille must be duct leakage to the outside. One variation of this method is to bring the ducts near the return grille to 50 Pa relative to the outside.
A number of potential errors are introduced using the flow hood method. First, the pressures from one leakage site to another are more variable than with the subtraction method. Second, not all the leakage from the ducts flows through the flow hood. Some of the leakage to outside flows through gaps around the registers and other communication locations between duct and house. This effect is the same as noted in the subtraction method. However this problem is much smaller since the open grille provides a preferred (lower resistance) flow path. This will result in lower leakage measurements than actually occur during the test.
The flow hood directly measures the flow through the open grille during these tests, rather than inferring it from two larger measurements as in the subtraction method. Pressures applied to the duct system with this testing method are usually lower than those applied by using the blower door subtraction or the Duct Blaster methods. The flow hood method measures flow more accurately, but due to restrictions within the duct system and duct leakage, a higher uncertainty about pressures is introduced. Traveling from the open grille, pressure is lost as restrictions or duct leakage are encountered.
The ability of this method to estimate leakage flow at a uniform test pressure is largely determined by how well the average pressure across leaks is estimated. If based on a series of pressure measurements, such as at a number of blocked grilles as well as at the plenums, the accuracy will improve. Once a determination of the leakage and pressure is made, the leakage at any other pressure can be estimated (See Estimation By Flow Exponent, p.30.)
The subtraction method and the flow hood method measure leakage at different pressures; therefore the results are not directly comparable. Figure 2 shows the relationship between the two tests when no corrections are made for leakage from the house to the ducts or for different pressures. The flow hood measurement measures higher leakage than the subtraction method on tight systems and lower leakage on loose systems. This data comes from tests conducted by Proctor Engineering Group on 42 houses.
Flow hood cfm50 =
134 + .45 2 Subtraction cfm50
Duct Blaster Method
The Duct Blaster measures the flow through the ducts to leaks both inside and outside the house (total duct leakage). Measurements are taken with the Duct Blaster attached at the blower compartment of the air handler or attached to the return grille. During these tests all registers are covered and the Duct Blaster flow is adjusted to create a reference pressure (usually 25 Pa) in the supply plenum or the nearest connected supply grille.
Potential errors using this method are more limited than the other two methods. One source of error continues to be the variability of pressures across the leakage sites. Other errors are operator error, location of the reference pressure probe, and variations in the seals at the register.
Pressure variations increase due to a restriction such as a coil, blower, or small duct work. Pressure variations are also effected by large leakage sites. When the Duct Blaster is installed at the blower compartment, the pressure variations across the leakage sites are less than with the flow hood because of any restriction in the return system. The blower compartment door is preferred because it reduces the possibility of restrictions or leaks in the return system from influencing the leakage readings. As with the flow hood, a series of pressure measurements at different locations in the duct system reduce the effect of pressure variation errors.
Operator error can be reduced by using digital time averaged measurements (of five seconds or longer), proper training and quality assurance, as well as step by step procedures.
The Duct Blaster measures the total duct leakage (leakage to inside plus leakage to outside). In order to determine the leakage from the ducts to outside the house, a house pressurization test has to be performed. The most common method of measuring the leakage to outside known as Blaster-blower door is to first bring the house to a specified pressure with the blower door. Then, by adjusting Duct Blaster flow, the reference location in the ducts is brought to zero pressure differential with respect to the house. If the pressure in the ducts is uniform, all the flow through the Duct Blaster is leakage to outside. Another method known as the Blasnik method, can also be used.
In a small series of tests, the Blasnik method and the Blaster-blower door methods of estimating leakage to the outside gave similar results.
Leakage Ratio Tests
A leakage ratio test provides a technician with a rapid method of estimating what portions of the leakage can be assigned to different areas. This may sound simple, but it is quite complex. Between floors, for example, is not necessarily inside the actual building pressure envelope. When tested with a blower door, basements may be more inside or more outside the pressure envelope. The effect of duct leakage in these spaces is only now under investigation (see The New Monster in the Basement p.37 and Stories from the Buffer Zone, p.40. )
The features of a good leakage ratio tool are the same as a good quantitative tool and the tool should be faster than measuring both of the leakages that make up the ratio.
The two primary ratio test methods are the Blasnik and the Half Nelson. The first quantifies the relationship between the leakage to inside the envelope and the leakage to outside the envelope. The second estimates the ratio of supply leakage areas to return leakage areas (see Table 2).
The Blasnik method is a valuable way of determining the proportion of duct leakage to outside versus inside. With the air handler fan off and with a blower door pressurizing the house to 50 Pa, two pressure readings are taken: pressure of the duct relative to the house (PD-H) and pressure of the duct relative to outside (PD-O). The ratio of the leakage between the duct and the house (QD-H) and the leakage between the duct and the outside (QD-O) when the duct is under pressure is computed (see Estimation by Flow Exponent, p.30).
This procedure is part of a method of estimating leakage flows without the use of a flow hood or Duct Blaster. The method adds a hole of known size to the duct system and by calculation estimates leakage.
Half Nelson--Supply-Return Split
Supply and return leaks have different impacts on energy use. The Half Nelson is a fast method which estimates the ratio between the total supply leakage area and the total return leakage area.
With all the registers sealed, the air handler is turned on and the pressures in the supply (PS) and return (PR) plenums are measured. The ratio of the total supply leakage area (AS) to the total return leakage area (AR) is estimated.
There are risks with this method. The test starves the blower motor for cooling air and should not be continued over a long period of time. It cannot be used immediately after repairs since the high pressures generated will blow out uncured mastic. In addition, John Tooley warns that these high pressures can damage duct board systems.
This procedure is part of a method of estimating leakage flows without the use of a flow hood, Duct Blaster, or blower door. The method (the Full Nelson), like the Blasnik method, adds a hole of known size to the duct system and calculates a leakage estimate.
Supply-return split can also be measured by conducting separate tests on both sections of the system with a blockage placed at the blower.
Qualitative measurements allow technicians to rapidly assess the areas of largest leakage and quickly check on progress of repairs. A good qualitative assessment tool
Provides clear and unambiguous direction for the technician.
Consumes as little time as possible to maximize technician time devoted to proper installation or repair of the distribution system.
The three primary qualitative methods are smoke stick, pressure pan, and register pressure (see Table 3.) Other qualitative methods include: tactile flow test, visual observation, and blocked return test.
Smoke Stick Method
Like the subtraction method, the smoke stick method is an extension of blower door technology. With the blower door pressurizing the house by 10-15 Pa, smoke released near a register will be more aggressively pulled into a register that has a major leak in that branch than a register that is distant from the larger leaks.
The pressure pan is a shallow pan (like a rectangular cake pan) that will cover and seal the supply or return register. The pan has a pressure tap that senses the pressure at the register when it is blocked off. Natural Florida Retrofit produces a combustion pressure pan and flow estimation device.
With the house pressurized to 50 Pa by the blower door, the technician records the pressure drop across the pressure pan when it blocks the register. If the pressure drop is less than half a pascal, any duct leaks are distant to that location. A larger pressure drop at one register (2-5 Pa) indicates that a large leakage site is near the location.
The pressure pan method is beneficial in prioritizing the attack on duct leakage sites, it can see leakage sites that are hidden in walls and under floors, and it provides a rapid check on progress (see Pressure Pan Takes the Cake, HE Mar/Apr '92, p.17).
Blocked Register Pressure
The blocked register test is an extension of the pressure pan technique, usable while the registers are taped shut. With the ducts pressurized by the Duct Blaster, the pressure drop across each taped register is measured by inserting a small probe. The register with the lowest pressure drop is near a large leakage site. If a few registers show low pressures relative to the remaining ones, it is likely that a significant leak exists near the branch of ducts. This method is less descriptive than the pressure pan.
Duct Diagnostic Decisions
The duct technician's tool box should contain a wide variety of diagnostic procedures to be used as conditions dictate.
Quantitative leakage measurement is best conducted with the Duct Blaster (and a blower door, if measurement of leakage to outside is required). We (with one exception) also consider the flow hood method useful.
The ratio methods are helpful since they estimate the leakage to particular areas. To check the integrity of the duct system if a blower door is in place, the pressure pan method is suggested.
We don't generally suggest the blower door subtraction method. It has the highest variability of the three quantitative methods described and provides weak feedback to the technicians sealing the duct system. The crew could be very successful at sealing the duct system, but would not see it indicated.
Figure 1. Duct leakage location categories.
Table 1. Quantitative Duct Leakage Tests
TEST Blower door subtraction Flow Hood Duct Blaster
EQUIPMENT Blower door Blower door and Combined duct pressurization
flow hood and flow device
ADVANTAGES * Inexpensive - only one * High certainty on * Inexpensive - only one piece of
piece of equipment required flow rate equipment required
* Good control over duct * Measures only leakage * Duct pressures well-controlled and
pressure to outside pressure distribution closest to that
* Measures only leakage in normal operating mode (except
to outside of envelope return is pressurized)
* Measures low flows accurately
* Measures total duct leakage
* Can be used on houses before drywall
COMMON Does not duplicate operating pressure distribution or flow (with the air handler fan on),
DISADVANTAGE resulting in under or over estimation of leakage at various points in the system.
DISADVANTAGES * Low repeatability under * Less control over * Requires a blower door to measure
windy conditions or on duct pressure leakage to outside
* Inaccurate for low flows * Requires two pieces * Overemphasizes leaks near registers
* Large piece of equipment of equipment to a lesser degree than the other
* Cannot test ducts before * Cannot test ducts two tests
drywall is installed before drywall is
* Overemphasizes leaks near installed
the registers * Overemphasizes leaks
near return registers
Figure 2. Idealized duct pressure distribution with fan on.
Home Energy * September/October 1993
Figure 3. Subtraction method results versus flow hood results.
Table 2. Leakage Ratio Tests
TEST Blasnik Half-Nelson
ESTIMATES Ratio: Duct leakage area Ratio: Supply duct leakage
to house / Duct leakage area / Return duct leakage
area to outside area
EQUIPMENT Blower door and Micromanometer
ADVANTAGES * Fast, once the registers * Inexpensive--only one
are sealed and blower piece of equipment required
* Can be used to convert * Fast, once the registers are
total leakage to leakage sealed
to inside and leakage * Can be used to convert
to outside system leakage to supply and
return leakages (which have
differing energy effects)
COMMON * Based on assumption that the flow exponent is .65 and duct
DISADVANTAGE test pressures are uniformly distributed
DISADVANTAGES * Should not be used on duct
board systems since they are
held together with tape
and high pressure will
ESTIMATION BY FLOW EXPONENT
Many of the newer duct diagnostics are based on a basic flow equation that relates the static pressure across the leaks to the air flow through the leaks. The relationship involves the pressure raised to a power--the flow exponent. This is a visual method of using the basic flow equation.
Three calculations referred to in Leak Detectors can be accomplished with this graph:
This graph is based on a flow exponent of .65 which is currently being used for duct leaks. Limited data exists to justify this exponent. Modera calculated average supply and return leakage flow exponents of .64 to .80 from Robison flow hood data. Duct Blaster testing by Proctor Engineering yielded exponents from .48 to .64. Large correction factors should be avoided unless the true exponent is known.
Correcting from one test pressure to another reference pressure.
Blasnik method for estimation of inside-outside split.
Half Nelson method for estimation of supply-return split.
Pressure correction. Estimates the leakage (Qr) at pressure (Pr) from the measured leakage (Qt) and test pressure (Pt).
Suppose the duct leakage (Qt) measured 200 cfm at a test pressure (Pt) of 12.5 Pa and we wish to estimate the leakage at a reference pressure (Pr) of 25 Pa.
With the registers sealed and the house pressurized to 50 Pa, pressure of the duct relative to the house (PD-H) was -16.7 Pa, the pressure of the duct relative to outside (PD-O) was 33.3 Pa, and we wish to estimate the ratio of the leakage between the duct and the house (QD-H) and the leakage between the duct and the outside (QD-O).
* Pressure ratio = absolute value of 33.3/-16.7 = 2
* Locate the Pressure ratio (2) on the left axis of the graph and trace that value across the gridline to the curve.
Supply-return split. Estimates the ratio of the total supply leakage area (AS) to the total return leakage area (AR).
Follow the gridline down from the curve to the bottom axis (Ratio B). Ratio B = 1.56
The duct leakage to inside is 1.56 times the leakage to outside.
With the registers sealed and the air handler on, the supply (PS) was 100 Pa, the return pressure (PR) was -200 Pa, and we wish to estimate the ratio of the of the total supply leakage area (AS) to the total return leakage area (AR).
* Pressure ratio = Absolute value of -200Pa/100 Pa = 2.
* Locate the pressure ratio (2) on the left axis of the graph and trace that value across the gridline to the curve.
* Follow the gridline down from the curve to the bottom axis (Ratio B). Ratio B = 1.56.
* The supply leakage area is 1.56 times the return leakage area.
Correcting from one Blasnik method for Half Nelson method
test pressure to another estimation of inside- for estimation of
reference pressure outside split supply/return Split
PRESSURE PR/Pt = PD-O/PD-H = PR/PS = RATIO Reference pressure Duct-outside pressure Return pressure ------------------ --------------------- --------------- Test pressure Duct-house pressure Supply pressure RATIO B Qr/Qt= QD-H/QD-O= AS/AR= Reference flow Duct-house flow Supply leakage area -------------- --------------- ------------------- Test flow Duct-outside flow Return leak. area RESULT Qr = Qt * B
Table 3. Qualitative Duct Leakage Tests
TEST Smoke stick Pressure pan register pressure
EQUIPMENT Smoke stick and Pressure pan and Duct Blaster
blower door blower door
ADVANTAGES * Fast once the * Fast once the * Fast once the
blower door is blower door is set up Duct Blaster is
set up set up and regis-
* Gives a numeric ters taped
reading that relates
to proximity and * Gives a numeric
size of leak reading that
relates to proxim-
* Can be completed ity and size of leak
without sealing registers
* Can be com-
* Can be a quick pleted without
stand in for removing tape
measured air flow from registers
DISADVANTAGES * Requires more * Once the registers * Requires that the
judgement (is are taped, it registers be taped
more ambiguous) requires removing
than the other the tape * Less descriptive
two methods than the Pressure
Discovering Ducts: An Introduction
Duct Fixing in America (Penn)
Duke Power's Success (Vigil)
Guidelines for Designing and Installing Tight Duct Systems (Stum)
Integrated Heating and Ventilation: Double Duty for Ducts (Jackson)
Managing Large-Scale Duct Programs (Downey)
Mobile Homes: Small Zones, Big Problems (Kinney)
New Group Hunts Bad Ducts (Obst)
The New Monster in the Basement (Treidler)
One Size Fits All: A Thermal Distribution Efficiency Standard (Modera)
Stories from the Buffer Zone (Kinney and Stiles)
Two Favorite Test Methods, By the Book (Modera)
Will Duct Repairs Reduce Cooling Load? (Parker, Cummings, and Meier)
Beauty and the Beast Upstairs (Legg)
Infiltration: Just ACH50 Divided by 20? (Meier)
Selecting an Infrared Imaging System (Snell)
Sizing Up Skylights (Warner)
Telecommuting: An Alternative Route to Work (Quaid)
User-Friendly Pressure Diagnostics (Fitzgerald, Nevitt, and Blasnik)
| Back to Contents Page | Home Energy Index | About Home Energy |
| Home Energy Home Page | Back Issues of Home Energy |
Home Energy can be reached at: firstname.lastname@example.org
Home Energy magazine -- Please read our Copyright Notice