12 Common Duct Design Mistakes

Most duct systems that end up noisy, short of air or expensive to run were not drawn badly; they were sized with shortcuts that look harmless on the plan. These are the twelve mistakes reviewers and commissioning engineers find again and again, with what to do instead.

The twelve mistakes

1. Sizing by velocity only

What happens: every duct gets a “reasonable” velocity and the fan is picked with a round number. Why it is wrong: the fan needs the total pressure of the index run, the path with the highest loss, which nobody calculated. Instead: size by equal friction at about 0.8–1.0 Pa/m (0.6 in quiet buildings), check velocities, then add up friction, fittings and terminals along the index run.

2. “10 % extra for fittings”

What happens: fittings are covered by a percentage on straight duct. Why it is wrong: fittings plus terminals are often more than half the total; in a worked office branch they were 48.6 Pa of 90.7 Pa. Instead: count every fitting with its loss coefficient ζ × velocity pressure (p_v = 0.6 v² Pa).

3. Mitred elbows without vanes

What happens: square elbows in the main because they are easy to fabricate. Why it is wrong: ζ ≈ 1.1–1.3, the same as 20–30 m of straight duct at 1 Pa/m. Instead: fit turning vanes (ζ 0.25–0.4) or use radius bends (r/D 1.5: ζ ≈ 0.15).

4. Fittings right at the fan

What happens: an elbow or a transition directly on the fan outlet or inlet to save plant room space. Why it is wrong: the uneven flow (system effect) robs the fan of pressure it shows on its curve. Instead: leave straight duct after the fan outlet as the manufacturer advises, or add the system effect loss to the calculation.

5. Flat rectangular ducts

What happens: 6:1 or 8:1 ducts to squeeze into a shallow ceiling. Why it is wrong: a 1000 × 200 mm duct has the same area as 500 × 400 mm but carries air like a 454 mm round duct instead of 488 mm, with a third more sheet metal (perimeter 2.4 m vs 1.8 m). Instead: keep aspect ratios at 3:1 or less (4:1 at most), prefer circular duct, and coordinate the void early.

6. Same-area round sizes for rectangular ducts

What happens: a rectangular duct is converted to the round duct with the same area. Why it is wrong: friction depends on shape; 500 × 400 mm has the area of a 505 mm round duct but the friction of a 488 mm one. Instead: use the Huebscher equivalent diameter D_e = 1.30 (ab)^0.625 / (a + b)^0.25.

7. Too much flexible duct

What happens: long, sagging or compressed flex runs to diffusers. Why it is wrong: its friction is several times that of rigid duct and rises sharply when compressed or kinked. Instead: limit flex to 1.5–2 m, fully extended and supported, with rigid duct up to the terminal.

8. No balancing dampers, or dampers nobody can reach

What happens: branches without dampers, or dampers above a fixed plasterboard ceiling. Why it is wrong: the system cannot be balanced and the near terminals take the air. Instead: show a damper on every branch, with access, and keep the excess pressure on short branches moderate.

9. Velocities too high near occupied rooms

What happens: ceiling mains at shaft velocities. Why it is wrong: regenerated noise rises steeply with velocity. Instead: typical office limits are 8–10 m/s in shafts, 5–7 m/s for ceiling mains, 4–5 m/s for branches and 2.5–3.5 m/s for runouts.

10. Leakage class not specified

What happens: “ductwork to be airtight”. Why it is wrong: for 300 m² of duct surface at 250 Pa carrying 2000 L/s, class A leaks about 14.7 % of the air, class C about 1.6 %. Instead: specify the pressure and leakage class, and a leakage test.

11. Cold ducts without a continuous vapour barrier

What happens: insulation with open joints, uninsulated hangers and dampers. Why it is wrong: a 13 °C duct in a 30 °C / 70 % space needs at least about 9 mm of insulation (λ 0.035) to stay dry, and every gap drips. Instead: check the thickness and specify a sealed, continuous vapour barrier.

12. Fire dampers and grease ducts done like ordinary duct

What happens: fire dampers installed outside their tested method or without access doors; kitchen exhaust drawn as normal galvanised duct. Why it is wrong: the damper’s fire rating is valid only as tested, and grease ducts need welded, liquid-tight construction with no fire dampers. Instead: follow the damper’s tested installation detail, show access doors, and design grease ducts separately (see kitchen hood exhaust airflow).

The full method, with a worked index run, is in our duct sizing guide; the airflows come from the loads in cooling load calculation basics.

Typical values — verify with the edition adopted locally.

FAQ

What is the most common duct design mistake?

Sizing ducts by velocity alone and never calculating the index run. The fan is then selected on a guess, so far terminals are short of air or the fan runs harder than it needs to.

How much pressure does a mitred duct elbow lose?

Without turning vanes, about 1.1–1.3 velocity pressures: at 6 m/s roughly 26 Pa, the same as 26 m of straight duct at 1 Pa/m. Vanes bring it down to about 6.5 Pa.

What aspect ratio is acceptable for rectangular ducts?

Keep it at 3:1 or less where you can, 4:1 at most. Flatter ducts use more metal, leak more and carry less air for the same area.