A noise specification or requirement should be more than just a value expressed in decibel. A full description of what is meant
is always necessary: is it about sound pressure level, and if that's the case is this a surface average value, what shape
of the measurement surface is used, and what the distance to the object? Maybe the sound power is used for a specification, but,
as is the case with many producers of refrigerators for example, the brochure doesn't mention that and the dealer doesn't know -
a careful (and expensive) effort of an engineer, killed at the desk of the marketing guys.
For a client we investigated the sound pressure level of a hydraulic generator package.
The Machinery directive requires determination of the highest sound pressure level at any bystander
position, e.g. at 1,6 m height and 1m distance.
For this package that sound pressure level was 85 db(A),
logically found close to a grid with a large fan behind it.
Consequently also the sound power level had to be determined. Therefor the surface average sound pressure level was measured and calculated. That appeared to be 77 dB(A),
also at 1 m distance. Note the difference of 7 dB with the sound pressure level at the bystander position, both measured at the same distance! For this client we made a Norsok compliant noise data sheet with clear description of all relevant values. The sound power level of this machine is 98 dB(A).
Complaints about the split unit of an airco led to investigation of the factory specs. These appeared to be confusing and incorrect.
The brochure and the technical documentation stated a "noise level" of 52 dB(A). Scrutinising the detailed specifications we found that the A-weighted sound pressure level of the unit was 56 dB(A).
The "noise level" value of 52 was a noise rating level and the (A) mentioned in the brochure and specifications did not apply to that rating level.
A case of misleading information caused by ignorance and/or carelessness. Noise rating is a value obtained by comparing a spectrum with a collection of reference curves. The rating level corresponds to the
highest curve touched by the curve of the measured spectrum. Only by chance a rating value can be close to LpA
This illustration explains the errors and the 4 dB confusion.
Noise measurement and specification is not a mystery, mysteries are created by carelessness and lack of knowledge.
We are eager to help avoid discussions and confusion, and costs!