How to Run a Total Duct Leakage Test at 25 Pascals

Why this matters

A duct system can look tight, carry mastic on every joint you can see, and still lose a fifth of its air through the joints you cannot: the panned return under the floor, the platform the air handler sits on, a boot gapped to the drywall. Energy codes and utility programs now ask for a number rather than a look, and the number is CFM25: how much air the ducts leak when held at 25 pascals. The test takes about an hour once you have done it a few times. It fails installs that looked finished, and the failures are almost always somewhere predictable. This article is the field sequence for the total-leakage version, which counts every leak regardless of where the air goes.

What total leakage counts, and what it does not

A total leakage test seals every register and grille, pressurizes the whole duct system to 25 Pa relative to the house, and measures the airflow the fan must supply to hold it there. Every cubic foot per minute the fan delivers is leaving through a hole somewhere, whether into the attic, the crawlspace, a wall cavity, or the living room.

It does not tell you where the air goes. A leak into a conditioned basement counts the same as a leak into a 130 F attic, even though one costs energy and the other mostly does not. Telling those apart is a leakage-to-outside test, which adds a blower door holding the house at the same 25 Pa so leaks to the inside stop flowing. Codes accept total leakage for new construction because it can be tested at rough-in, before drywall, when sealing is cheap. California has its own test procedures and acceptance values (See related: California Title 24 Duct Tightness Test SOP).

25 Pa is a test convention, a pressure in the range residential ducts commonly see in operation. It is a test condition, not an operating pressure: the real system runs some sections higher and some lower, which is why the result is a comparison number, not a prediction of exact leakage in service.

Start state

Hazards, each with its control. You will be in an attic, crawlspace or closet: boards across joists in the attic, a light and a second person aware you are in a crawlspace. The air handler is turned off at its service switch, locked and tagged, before you open its cabinet to connect the fan, and proven dead with a tester proven on a known live source before and after, per NFPA 70E Article 120 (120.5 in the 2021 edition, renumbered in 2024), with work practices at 29 CFR 1910.333(b)(2). On a gas furnace, the gas valve is off as well, since the burner compartment shares the cabinet you are opening. Ladders to ceiling registers are set on a stable footing; most register sealing is done from a ladder.

What you need: a calibrated duct pressurization fan with its flow rings, a two-channel digital manometer, tubing, register-sealing film or tape that will not lift ceiling paint, a cardboard or plastic transition to fit the fan to the return or the cabinet, and the conditioned floor area of the house from the plans.

Check before you start: the filter is out (a loaded filter between fan and ducts changes nothing about the leaks but makes the fan work harder and can hide a slot leak behind it), every supply and return is accounted for on the plans, and the dampers on any zoning system are driven open.

The setup

  HOUSE (reference for channel A)

  [film] supply  [film] supply  [film] supply
     |              |              |
  ===+==============+====+=========+=== SUPPLY TRUNK
                         |
              tube A --> + supply boot nearest
                         |  the air handler
                  [ AIR HANDLER ] blower off,
                         |        filter out
                    return plenum
                         |
  [film] return -- [ TEST FAN ] fitted at the
                      ring       central return
                     tube B --> fan pressure
                                 = flow

Channel A reads duct pressure relative to the house, through a tube pushed into a supply boot near the air handler, out of the fan's direct blast. Channel B reads the fan's own pressure, which the manometer converts to airflow for the ring installed. Every other register and grille is sealed with film.

The steps

1. Seal every register and grille except the one the fan uses. Cover each one completely with film pressed down at its edges. Skip one (the forgotten laundry room supply, the return behind a couch) and the test measures that open register as leakage, then the house fails for a hole that is not a leak. Count the sealed openings against the plans.

2. Fit the fan. At the central return grille, or directly at the air handler's blower compartment with a transition, sealed so the fan's only path is into the ducts. A leaky fan connection reads as duct leakage.

3. Connect and zero the manometer. Tube A into the supply boot, tube B to the fan's pressure tap, reference ports open to the house. Set the manometer to the fan model and ring you installed. Why the ring matters: the conversion from fan pressure to flow is different for each ring, and the wrong setting produces a confident wrong number.

4. Bring the duct to 25 Pa. Increase fan speed until channel A reads 25 Pa duct to house. Use the manometer's averaging for at least a few seconds; wind gusts and a door closing elsewhere in the house move the reading.

5. Read and record CFM25. Read channel B as flow. Record it with the ring used, the duct pressure actually held, and the time. If the fan reading is low on the ring's range, step down to a smaller ring and re-read; a fan running near the bottom of its ring range is less accurate.

6. Normalize and compare to the limit. Most programs and codes express the limit per 100 square feet of conditioned floor area. IECC Section R403.3 (the duct testing and leakage provisions) in the 2021 edition sets total leakage at no more than 4.0 CFM25 per 100 sq ft for the postconstruction test, with different values for rough-in and for ducts fully inside the thermal envelope; it applies where your jurisdiction has adopted that edition, and local amendments and other editions differ. Utility programs may use their own limits. Use the one that governs this job.

7. If it fails, find the leaks with the fan running. Hold 25 Pa and work the system with a smoke puffer or a damp hand along every joint you can reach. The common places are listed in the example below.

8. Restore everything. Fan off and removed, film off every register and grille, filter back in, the cabinet panel on, gas back on at the valve, lock and tag removed, power restored. Run the system and confirm airflow at the registers, including the ones you had to climb to. Skip the walk-through and a single film left on a bedroom supply becomes a cold-room callback the next day.

Worked example: a new install that failed at 8.1

A new 3-ton system in a 2,000 sq ft single-story house, attic ductwork, air handler in a hall closet on a framed return platform. The governing limit is the 2021 IECC postconstruction value as adopted here: 4.0 CFM25 per 100 sq ft.

The limit for this house: 2,000 / 100 = 20 hundreds of square feet, times 4.0 = 80 CFM25 allowed.

First test. Twelve supplies and two returns sealed, fan at the central return on its middle ring, duct pressure held at 25 Pa: 162 CFM25. Normalized, 162 / 20 = 8.1 CFM25 per 100 sq ft, about double the limit. Checked the other direction: 80 CFM25 allowed is 162 x 0.49, so the house needs to lose about half its measured leakage.

Where it was. With the fan holding 25 Pa, smoke at the closet found the big one: the framed return platform under the air handler had no sealant at its drywall seams, and with the ducts pressurized, smoke streamed out of those seams into the closet and on up past the closet's top plate into the attic. In the attic, two boots had no seal to the drywall, and a supply trunk end cap had tape lifting at one corner. The flex joints, all mastic-sealed, showed nothing.

What the crew did. Mastic and fabric over the platform seams and the cabinet-to-platform joint, the two boots sealed to drywall from above, the end cap re-sealed with mastic.

Second test. Same ring, same setup: 71 CFM25. Normalized, 71 / 20 = 3.55 CFM25 per 100 sq ft, under 4.0. Pass, with 80 - 71 = 9 CFM25 of margin.

What the numbers say about the job. The sealing took out 162 - 71 = 91 CFM25, and by the crew's own observation most of it was the platform, a joint the duct installer considered part of the carpentry and the carpenter considered part of the ductwork. The ticket records both tests, the ring, the leak locations, the sealing done, and the final value against the limit.

When the fan cannot reach 25 Pa

A very leaky system may not reach 25 Pa even at full fan speed on the open ring. That is already a result: the system is far from passing, and sealing comes before testing. Most programs allow a reading at the highest pressure achieved, corrected to 25 Pa by a formula the program states, and some do not accept it at all. Record the pressure you reached, the flow at that pressure, and which correction you applied, so the next tester can repeat it.

The opposite case is a system so tight the fan cannot be slowed enough to read at 25 Pa on its smallest ring. That is a pass by any limit, but record the reading the way the program wants, since some want a value, not a description.

How to verify you got it right

• Every register and grille is counted against the plans and sealed, and every film is off again with airflow confirmed at each one afterward.
• The recorded CFM25 carries its ring, the duct pressure held, and the floor area used to normalize it.
• The normalized result is compared against the limit that actually governs this job, with its edition or program named on the record.
• Air handler power and gas are restored, the filter is back in, and the system runs.

References

• IECC Section R403.3 (2021 edition), duct testing and leakage, as adopted by your jurisdiction, with local amendments
• NFPA 70E, Standard for Electrical Safety in the Workplace, current edition, Article 120; 29 CFR 1910.333
• Duct pressurization fan and manometer manufacturer instructions: ring ranges, flow conversion, low-pressure correction
• See related: California Title 24 Duct Tightness Test SOP; How to Install a Takeoff and Flex Branch to a New Register