What is A-weighting?


A-weighting, defined by IEC 61672-1:2013, was established to take into account the average sensitivity of human’s ear as a function of frequency. A-weighting is used to convert a physical quantity of acoustic pressure (in dB) in a psychoacoustic quantity (in dBA) that is supposed to better quantify how loud a noise is perceived by humans.

The following figure shows a comparison between SPL without weighting and SPL with A-weighting (noted ASPL in MANATEE software) as well as the A-weighting curve. One can see that SPL magnitude at lower frequencies is reduced due to to A-weighting curve.

Sound power level with a-weighting and without weighting comparison
Sound power level with a-weighting and without weighting comparison

Application to e-NVH

As the ear is less sensitive to low frequencies (< 100 Hz), the A-weighting gives less weight to that part of the acoustic energy spectrum. The part of the spectrum near 2500 Hz has the highest sensitivity and electrical systems often whine in this frequency range, that’s why the electromagnetically-excited noise of rotating machines can be so annoying.

As acoustic noise due to Pulse Width Modulation mainly occurs at once or twice the switching frequency, some of the worst choise of switching frequency are 1250 and 2500 Hz. For induction machines where noise often occurs at the rotor slot passing frequency, one should also avoid having ZrN/60 close to 2500 Hz (Zr: rotor slot number, N: nominal speed in rpm).

Application to MANATEE

MANATEE software allows to take into account this A-weighting factor on Sound Pressure Level or Sound Power Level with several post processing such as plot_ASPL_fft, plot_ASWL_13oct, plot_ASWL_log.

The A-weighting acoustic noise measurement is stated in IEC 60034-9 for noise limits of rotating electrical machines.

More complex sound quality metrics might be used to refine the subjective assessment of sound perception (loudness, roughness, sharpness).

Previous Next