Discussion about this post

User's avatar
George's avatar

Graham, you're onto a real issue, and you're right on one key point: standard A-weighted (dB(A)) limits are a poor way to judge this kind of noise, because they deliberately down-weight low-frequency energy, so a data centre can sit inside its dB(A) limit while the hum is still intrusive indoors. Where the piece goes wrong is the framing around infrasound. True infrasound sits below about 20 Hz; but the drone people actually describe near these sites, the "refrigerator" hum off the cooling plant, fans and transformers, lives mostly in the low-frequency band above that, roughly 50–200 Hz, as it is based on the grid frequency harmonics. That is audible, and it is entirely measurable. The problem isn't that the instruments can't detect it; it's that dB(A) under-represents it. Measured properly, with C-weighting (an LC–LA gap above about 15–20 dB is a reliable red flag), one-third-octave band analysis, an indoor low-frequency assessment and a tonal penalty, low-frequency noise can be characterised, conditioned and enforced at the planning stage. The fix is a better metric and better technology, not the conclusion that it can't be measured or regulated.

4 more comments...

No posts

Ready for more?