Why the test pressure has to travel with the number
Air through a hole does not flow in proportion to pressure. It follows a power law: double the pressure and the flow goes up by two raised to the exponent, which for a duct system is typically somewhere around 0.6. Two to the power 0.6 is about 1.52, so the same duct system that leaks 96 CFM at 25 pascals leaks around 146 at 50.
That is a fifty percent difference in the headline number with nothing changed about the duct. A leakage figure quoted without its pressure is therefore not a measurement, it is a rumour, and comparing two systems tested at different pressures without converting is comparing nothing at all.
The exponent, and why this page asks for it
The exponent describes the shape of the leaks. A clean sharp-edged hole gives roughly 0.5, because flow through an orifice goes with the square root of pressure. A long narrow crack gives something closer to 1.0, because viscous flow through a thin passage is nearer linear in pressure. Real duct systems are a mixture of both — panned joists, taped seams that half let go, boots that never got mastic — so they land in between.
You can measure it, if the test equipment takes readings at several pressures: the exponent is the slope of the log of flow against the log of pressure. Most single-point tests do not, so the exponent gets assumed. This page makes the assumption visible rather than burying it in a conversion, because it is the single largest source of error whenever a reading is converted between pressures.
| Exponent | 96 CFM at 25 Pa becomes, at 50 Pa | And at 10 Pa |
|---|---|---|
| 0.50 | 135.8 | 60.7 |
| 0.60 | 145.5 | 55.4 |
| 0.65 | 150.6 | 52.9 |
Between the top and bottom rows there is about eleven percent, on a conversion across one doubling of pressure. Convert further and it grows.
Three ways of normalising, three different questions
Leakage per 100 square feet of floor area is the form most codes and programmes use, because it scales with the size of the building and can be written as a single limit. Leakage as a share of system airflow answers a different question — how much of what the blower moves never reaches a room. Leakage per ton is a third framing, closer to the equipment than to the building.
They do not rank systems the same way. A tight system on a large house can be well inside a per-area limit while losing a substantial share of its airflow, if the blower is small. The page reports all three because a service record that only has one of them cannot be re-read later for the other question.
Total leakage and leakage to outside
These are two different tests and the difference is often large. Total leakage measures everything the duct system loses, including seams inside the conditioned space that leak into a wall cavity that is itself inside the building. Leakage to outside measures only what escapes the thermal boundary, and the test is set up so that leaks into conditioned space do not register.
The energy consequence is completely different. Air lost into a room that is inside the boundary is distribution waste — it heats the wrong room rather than leaving the house. Air lost into a vented attic is conditioned air gone, plus an equal volume of attic air pulled in somewhere else to replace it. Which test was run is therefore as important as which pressure it was run at, and the selector at the top of the page carries it into the result and the share text.
Where the rest of the tightness picture is
The building shell version of this same test, where a blower door reading at a test pressure is converted into an equivalent hole area, is the blower door leak area calculator — the same power law and the same exponent question, applied to the envelope instead of the duct. What shell tightening is worth in fuel is the air sealing payback calculator, and the wider picture is in the home energy audit guide.
On the duct side, the pressure the system runs at in normal operation rather than under test is the external static pressure calculator, and the airflow it is meant to be distributing is in the duct size calculator and the duct branch balancing calculator.
Questions people ask
What is the maximum duct leakage allowed?
That is set by whichever code or programme has been adopted where the work is, and it changes between jurisdictions and between code cycles. It also depends on which test was run and at what stage of construction. This page therefore takes the limit as an input and restates it beside your measurement rather than stating one, because a number published as a fact here and applied under a different adopted code would be worse than no number. The authority that adopted the limit decides whether a result passes.
What exponent should I use if I do not know mine?
If the test was single-point you do not know it, and the honest thing is to note which value you assumed alongside the result. Duct systems generally fall between 0.5, the value for flow through clean sharp-edged holes, and 1.0, the value for viscous flow through long narrow cracks. The table in the guide above shows what the choice is worth: converting 96 CFM from 25 to 50 pascals spans about eleven percent across a plausible range of exponents. The best answer is to run the test at the pressure you have to report at, so no conversion is needed.
Why does 96 CFM at 25 Pa not become 192 at 50 Pa?
Because flow through a leak is not proportional to pressure. It follows pressure raised to the exponent, which for duct systems is well below 1. At an exponent of 0.6, doubling the pressure multiplies the flow by two to the power 0.6, or about 1.52 — so 96 becomes about 146, not 192. If the flow really did double with pressure, the exponent would be 1.0 and every leak in the system would be a long thin crack.
Should I test total leakage or leakage to outside?
Whichever the procedure you are reporting under calls for, and they are not interchangeable. Total leakage counts everything the duct loses; leakage to outside counts only what crosses the thermal boundary, and on a system with all its ductwork inside conditioned space the second number can be a small fraction of the first. Record which one was run — a leakage figure without both its pressure and its test type attached cannot be compared with anything later.
How much air is that in real terms?
The share-of-system-airflow line is the one to look at for that. At the defaults here, 96 CFM against 1,200 CFM of system airflow is 8 percent of everything the blower moves. Whether that matters depends on where it goes: inside the conditioned space it is distribution waste that heats the wrong part of the house, and into a vented attic it is conditioned air leaving the building with an equal volume of attic air drawn in to replace it. The same number, two very different consequences.