In the current environment of noise abatement, high frequency noises are especially objectionable. For gear drives, high frequency noises are symptoms of gear mesh-frequency vibrations which are ever-present due to the imperfect conjugate action between the gears. These vibrations are then transferred to the gear housing to be emitted as noise. The vibration energy flow can be disrupted, for example using a hydrostatic bearing as a low pass filter of vibrations while still preserving the positioning accuracy of the gear drive.
Hydrostatic bearings are a class of fluid bearings consisting of externally pressurized fluid trapped in a recess to provide the necessary bearing support. Previous research has assumed the incompressibility of the working fluid in a hydrostatic bearing. In practice, the working fluid is not incompressible and the effect is not necessarily insignificant. Using the compressibility of the working fluid and the unique properties of a hydrostatic bearing, a low-pass filtering effect on the gear mesh-frequency vibrations can generated. This low-pass filter allows low frequency vibrations to be transmitted while high frequency vibrations are attenuated. The vibration attenuation disrupts its transmission to the gear housing and prevents its conversion into high frequency acoustic noise.
Models to predict this low-pass filtering effect were developed using a control volume approach for constant flow and capillary flow compensation schemes while the compressibility of the working fluid is incorporated using the bulk modulus. A linear differential equation was then obtained. Using the solution to the differential equation and convolution, an expression for the frequency response of the transmitted dynamic pressure was developed. In addition, a proof of concept experiment was designed and implemented to validate the theoretical models. The experimental data shows similar attenuation of the transmitted dynamic pressure according to theoretical predictions for both the constant flow and capillary compensated flow models.