Technical note

Blower noise, and the fifth power hiding inside it

Total fan sound power varies roughly as the fifth power of tip speed. That single fact explains why a modest speed increase can breach a boundary limit, and why a bigger, slower machine is often the cheapest silencer there is.

The rule

The fifth power

Raise the logarithm and the fifth power becomes an easy sentence: the change in sound power level is fifty times the log of the tip speed ratio. Ten per cent more speed is about 2 dB. Double the speed is about 15 dB, which is not a small number in a residential boundary assessment.

Speed ratio 1.1
50 x log(1.1) = about 2 dB louder
Speed ratio 1.25
about 5 dB
Speed ratio 1.5
about 9 dB
Speed ratio 2
about 15 dB
Density ratio
20 x log of the density ratio, so light hot gas is quieter than cold air at the same tip speed

Tip speed is the product of diameter and rpm, so the same rule is the reason a larger wheel turning more slowly is the standard engineering answer to a noise problem. It is also the reason that a 15 per cent flow increase, which costs 50 per cent in power, costs about 9 dB – and that 9 dB is what the neighbours notice.

Vocabulary

Sound power and sound pressure are different

Sound power, L_W

What the fan emits, a property of the machine, referenced to one picowatt. Quoted in octave bands. This is the number a manufacturer should give you, because it does not depend on where you stand.

Sound pressure, L_P

What your ear or a meter measures, referenced to 20 micropascals. It depends on distance, directivity, reflections and the room. A site measurement is a sound pressure figure, and comparing it to a manufacturer sound power figure without converting is how a compliant fan gets accused.

Which spectrum

A dBA figure tells you how it will be judged against a limit. It does not tell you whether a tonal blade-pass component is hiding inside it. Specify the octave band data if a boundary limit or a room acoustic is in play, and ask for the measurement method – the laboratory sound method for fans is the companion to the aerodynamic test method, not part of it.

Control

The control list, cheapest first

Slow it down and grow it

The fifth power works in reverse. A frame size larger at lower rpm is usually quieter, cheaper to run and easier on bearings.

Fix the inlet

An unobstructed, properly transitioned inlet is free. Elbows bolted to a bellmouth are a classic source of regeneration noise, and they cost nothing to avoid.

Isolate it

Flexible connections and an inertia base stop structure-borne noise going into the building. Silencers deal with the airborne path; nothing here deals with a duct that is ringing like a bell.

Then attenuate

Inline or circular silencers, sized for pressure loss so the fan is reselected with the silencer in the system. A silencer added to an existing selection is an unmeasured pressure drop.

Questions

Noise questions

Two identical fans side by side: how much louder?

About 3 dB in sound power, and near-field behaviour between them is worse than the arithmetic suggests. Two machines running on one system is a selection question, not an addition.

Is a backward curved fan quieter than a radial one at the same duty?

Usually yes, at the same duty and the same tip speed, because less of the energy leaves as turbulence at the blade exit. Tip speed is the controlling variable far more often than wheel type.

How do I compare a dB figure to a noise limit at the boundary?

The limit is in sound pressure at a distance and a height, with a stated background and time period. The manufacturer figure is sound power. The conversion needs distance, directivity and the room or terrain – a straightforward calculation, but not something you can do by subtracting a fixed number.

Next

Related notes

Keep reading

Call the works Enquire