AU08 (13.12.2016 19:08 - 13.12.2016 19:19, файл(ы), ссылка) , в ответ на Только надо понять, не было ли там коэффициента усиления для глюков в капаситоре...Он поди на выходе усилка мерил... автор: Evgeny_CD
В конце комментариев ссылка на другую статью "Reducing MLCCs' piezoelectric effects and audible noise"
A MLCC alone is in most cases not sufficient to generate a problematic or disruptive Sound Pressure Level (SPL).
But soldered on a PCB board the MLCC generates a spring mass system, which increases or dampens the oscillations depending on the frequencies (Figure 1).
However, while a lower rated capacitor usually exhibits a higher capacitance at a given voltage than a higher rated one, the latter will tend to generate more noise.
Different contact area (width, length) will have little impact on the noise created but for the same capacitance a thick capacitor will generate a lower SPL than a thinner one. It can be noted that the thinner the capacitor the higher the electric field (as the layers are closer) and the higher the bias effect. For example, a 1 mm capacitor will produce 13 dB more than a 2.5 mm one.
The self-resonating frequency of the board being as difficult to measure as to control, it can be ignored. Generally speaking the thicker the board, the less entitled to deformation it is, and thus the lower SPL it produces. At the same time, the more important the board surface surrounding the vibrating capacitor, the louder will be the noise. A capacitor placed at the edge of the PCB will be preferred. Measurements showed that a reduction from 2 mm to 1 mm thickness worsened 5 dB while a reduction from 14 cm² to 5 cm² improved 6 dB.
Placed next to each other, capacitors generate higher overall SPL (+14 dB between a single capacitor and three placed in parallel).