Generator, chiller and rooftop plant noise — meeting the night limit, and proving it
Every large building has machinery that has to run and neighbours who have to sleep. The standby generator tested every week, the chillers and cooling towers running all night, the extract fans, the water pumps, the kitchen exhaust. Building services are the leading source of complaints against condominiums, hospitals, shopping centres and office buildings — and the thing the equipment vendor guarantees with a single number that rarely matches the installed condition.
Geonoise assesses, predicts, measures and verifies building-services noise from before the purchase order to after installation, so that the level at the boundary and in the neighbouring bedroom meets the night-time limit, and so the owner has evidence of how many decibels the silencer, the enclosure or the isolated plinth actually delivered. We do not sell the equipment we recommend.
Where we are engaged
- Complaints from neighbouring or upper-floor residents about a generator, chiller or fan — particularly at night and during test runs
- A letter from the district office, municipality or public-health authority requiring measurement and remediation of nuisance noise
- Before ordering new plant — predicting the level at the boundary and in the nearest rooms from the manufacturer's sound power, to set the silencer specification and the position while they can still be changed
- After installation — verifying that the plant and its silencers achieve what the vendor guaranteed, before the retention is released
- Generator test regimes — setting the time, duration and load of the weekly test so that it complies and can be notified to neighbours in advance
- Hospitals and buildings with several generator sets that must be tested under real load and that draw complaints from wards or the community
The limits that have to be met
Building services are judged against three layers of criteria, and passing one does not mean passing the others.
- General level at the boundary or receiver — Leq 24 h not above 70 dBA and Lmax not above 115 dBA, under National Environment Board Notification No. 15 (B.E. 2540)
- Nuisance level — the difference between the level with the source running and the background not above 10 dBA, under Notification No. 29 (B.E. 2550) and the B.E. 2565 nuisance-noise notification. This is the criterion a chiller fails most often, because the background at two in the morning is very low
- Internal criteria — background noise in bedrooms, wards or offices against the NC or NR value the building was designed to, and structure-borne noise from vibrating plant on the roof or a plant floor
What we do
- Measure at the boundary, in the neighbouring rooms and in the plant room under real operating conditions, including night-time and generator tests, with calibrated Class 1 instruments.
- Predict levels from the manufacturer's sound power in a SoundPLAN model to ISO 9613-2, to compare positions, orientations and mitigation options on paper before anything is built.
- Measure the plant's actual in-situ sound power to ISO 3747, to compare against the vendor's claim.
- Measure the actual insertion loss of silencers in situ to ISO 11820 — catalogue values are obtained in a test rig with no bends, no real flow velocity and no breakout through the casing.
- Rank the sources with the acoustic camera where the generator, cooling tower and fans share one roof, so that the one governing the level at the complaint position is known.
- Measure vibration at the plant base and in the structure, for noise that travels down the building rather than out through the air.
- Specify verifiable measures — silencers, enclosures, screens, isolated plinths — with a target level at the receiver, and re-measure after installation.
The vendor's guarantee, and what it does not cover
A generator is offered with one number, say 75 dBA at 7 m. That figure was measured in an open field, at a load the manufacturer chose, without the exhaust run actually fitted, the plant-room louvres or the neighbouring wall that reflects the sound. At the real complaint position the level can differ substantially, in either direction. Before signing we read what the acoustic guarantee actually promises; after installation we measure whether it was met, while the retention is still held. See Acoustic Contract & Guarantee Review.
What you receive
- A measurement report at the boundary and receiver positions against all three layers of criteria, with calibration certificates, suitable for submission to the authority
- Source ranking and each machine's contribution at the complaint position
- Predicted levels for new plant or new positions, with options compared
- A verifiable specification for the measures, with target levels and the acceptance method, and no stake in any vendor
- Post-installation verification, before and after, to close the complaint and release the retention
- Where wanted, a generator test regime and method that complies and can be notified to neighbours
Standards & method
National Environment Board Notifications No. 15 (B.E. 2540) and No. 29 (B.E. 2550) · Nuisance-noise notification (B.E. 2565) · ISO 1996-2 · ISO 9613-2 · ISO 3747 (in-situ sound power) · ISO 11820 (silencers in situ) · ISO 16032 (building services noise) · NC / NR criteria · IEC 61672-1 Class 1 · instruments calibrated by an ISO/IEC 17025 accredited laboratory
FAQ
Our generator only tests for 30 minutes a week. Why are we getting complaints?
Because the nuisance criterion uses the difference from the background at that moment, not a daily average. A test in the early morning or at night, when the background is low, exceeds 10 dBA easily. Choosing the test time, the load and advance notification is part of the answer, not just a silencer.
The vendor guarantees 75 dBA at 7 m. Can we rely on that?
Under the conditions it was measured in — usually an open field, a load the manufacturer chose, and no real exhaust run or plant room. At your complaint position the figure can differ substantially. We predict from the sound power before ordering and measure after installation against the guarantee.
The chiller complies by day but not at night. Why?
The nuisance criterion is measured against the background, which at two in the morning can be 15 dBA lower than by day. The same machine at the same level therefore becomes an exceedance. Remedies include quieting the plant, reducing speed at night, and repositioning or reorienting — costs to be compared on paper first.
We fitted a silencer and it is still loud. Why?
Usually because the catalogue insertion loss was measured in a test rig, while the real installation has bends, high flow velocity and noise breaking out through the casing and the machine itself rather than leaving through the end of the duct. We measure the actual insertion loss in situ to ISO 11820 and use the acoustic camera to find where the sound is escaping.
How do you tell airborne from structure-borne noise?
By measuring vibration at the plant base and in the structure of the complaining room, together with the spectrum in that room. Noise travelling down the building is fixed with vibration isolation, not a silencer — and choosing the wrong one means paying twice.
Is there anything special about hospitals?
Yes: wards with low background-noise criteria, several generator sets that must be tested under real load, and notifications that have to be coordinated with clinical staff. We design the test and measurement method around those constraints.
Do you need the plant switched off to measure?
The main measurement is done under normal operation. Short switch-off tests, one machine at a time, separate the contributions most clearly and are coordinated with building management to minimise disruption. Where switching off is impossible, the acoustic camera does part of that job.
Do you have a conflict if we ask you to fix it?
No. We do not sell or install silencers, enclosures or isolation plinths. We write a verifiable specification you can procure from anyone, then re-measure to confirm it delivered what was promised.