Duct velocity chart
The velocity bands duct systems are actually designed to, set out by application and by position in the system — main, branch, riser and outlet — in both metres per second and feet per minute, with the noise reasoning behind each band and what happens either side of it.
Recommended velocities by position in the system
| Application | Main duct | Branch | Runout to outlet |
|---|---|---|---|
| Residential | 3–5 m/s 591–984 fpm |
2.5–4 m/s 492–787 fpm |
1.5–3 m/s 295–591 fpm |
| Offices, libraries, consulting rooms | 4–6 m/s 787–1181 fpm |
3–4.5 m/s 591–886 fpm |
1.5–3 m/s 295–591 fpm |
| Hospitals, studios, theatres | 3.5–5 m/s 689–984 fpm |
2.5–4 m/s 492–787 fpm |
1.5–2.5 m/s 295–492 fpm |
| General commercial, retail | 5–8 m/s 984–1575 fpm |
3.5–5.5 m/s 689–1083 fpm |
2–4 m/s 394–787 fpm |
| Schools, restaurants, banks | 4.5–6.5 m/s 886–1280 fpm |
3–5 m/s 591–984 fpm |
2–3.5 m/s 394–689 fpm |
| Industrial, plant rooms | 8–12 m/s 1575–2362 fpm |
6–9 m/s 1181–1772 fpm |
4–6 m/s 787–1181 fpm |
Conventional design bands for low-velocity systems. Where a project specification or a room acoustic criterion names a limit, that number governs over any table.
The velocity ladder, and why it descends
Read any row of that table from left to right and the velocity falls by roughly half between the main and the runout. That descent is deliberate, and it is a noise strategy rather than a sizing one.
Air moving in a straight duct is remarkably quiet. What generates sound is turbulence at discontinuities — take-offs, elbows, dampers, transitions — and the sound power those fittings produce rises with something close to the fifth or sixth power of velocity. Halving the velocity at a fitting can drop its sound power by 15 decibels or more. In a main duct that hardly matters, because there is length, bends and often lining between the fitting and the nearest ear. At a runout take-off a metre above a ceiling tile, there is nothing. So the system is designed to spend its velocity where the building can absorb the noise, and to arrive at the diffuser slowly enough that the diffuser itself sets the sound level.
Main duct bands used by the calculators
| Application | m/s | fpm |
|---|---|---|
| Residential | 3–5 | 591–984 |
| Commercial — low noise | 4–6 | 787–1181 |
| Commercial — general | 5–8 | 984–1575 |
| Industrial | 8–12 | 1575–2362 |
| Branch / runout | 2–4 | 394–787 |
These are the bands the ductulator and the duct velocity calculator check their results against, generated from the same data so the pages cannot drift apart.
What the velocity number does not cover
- Fittings, not straight duct, make the noise. A run at the top of its band with generous radius bends can be quieter than a run in the middle of its band with a hard tee and a throttled damper.
- A throttled damper is a noise source. Balancing a badly distributed system by closing dampers puts a high-velocity restriction exactly where you least want one. Size to balance, then trim.
- Flexible duct changes everything. Its roughness generates more turbulence at the same velocity, and a compressed flex connection close to a diffuser is one of the most common causes of complaints in an otherwise sound design.
- Return needs its own check. Return grilles often sit closer to occupants than supply diffusers, with less attenuation in the path.
- Extract systems may run high on purpose. Kitchen, fume and dust extract need transport velocity to stop deposition, and that requirement outranks the comfort bands entirely.
Frequently asked questions
What is the recommended air velocity in a duct?
For a main duct in general commercial work, 5 to 8 metres per second — about 1000 to 1600 feet per minute. Where noise matters, offices and hospitals sit lower at 3.5 to 6 m/s. Branches run below their main, and final runouts to a diffuser drop to 1.5 to 4 m/s so the terminal rather than the duct sets the sound level. Industrial systems run 8 to 12 m/s and higher where particulate transport demands it.
What happens if duct velocity is too high?
Three things, in order of how quickly you notice them. Noise comes first, and it comes from fittings rather than straight duct — take-offs, elbows and dampers convert air speed into sound far more efficiently than a straight run. Then fan power, which rises steeply because pressure loss scales with velocity squared. Then, at the extreme, erosion of duct sealant and lining, and in flexible duct, movement of the duct itself.
What happens if duct velocity is too low?
You pay for it in metal, void space and coordination rather than in noise or energy. A very low velocity means a large duct, which is more sheet, more insulation, more hangers and a deeper ceiling void — and in a building where floor-to-floor height is fixed, void depth is expensive. Low velocity also makes balancing harder, because the pressure differences the dampers work with become small.
Why do branch ducts run slower than mains?
Because branches are closer to occupied space and closer to fittings. Sound generated in a main has duct length, bends and lining to attenuate it before it reaches a diffuser; sound generated at a runout take-off has almost nothing between it and the room. The velocity reduction going down the system is a noise strategy, not an accident of sizing.
Do these velocities apply to return and extract duct?
Broadly yes, and often return is designed slightly lower than supply for the same space because return grilles are frequently closer to occupants and have less attenuation available. Kitchen and fume extract are the exception in the other direction: they run high deliberately, because grease-laden and contaminated air has to be kept moving to stop it depositing in the duct.
How does velocity relate to friction rate?
They are two constraints on the same decision, and either can govern. Sizing on equal friction alone can produce a duct that meets the friction target but runs too fast for where it is; sizing on velocity alone can produce one that is quiet but has an unacceptable pressure loss over a long run. Good practice applies both and lets whichever needs the larger duct win — which is exactly what the ductulator does, and it tells you which one governed.
Last updated: 26 July 2026