An experimental study consisting of three separate investigations is conducted on the system-level behaviors of a planetary gear set. A test rig is designed and procured for the purpose of measuring (i) transmission error, (ii) planet-to-planet load sharing and (iii) sun gear radial orbits under quasi-static conditions within a range of input torque. Several test matrices are implemented and are designed to quantify the effects of gear manufacturing errors and modifications as well as kinematic configurations under which the planetary gear set operates in a repeatable and accurate manner.
The test matrix for the planetary gear set transmission error study includes three distinct phasing conditions (in phase, sequentially phased and counter-phased) of a four-planet gear set as well as two planet tooth profile modifications and two sun gear run-out error values. Two distinct power flow conditions with a fixed planet carrier and a fixed ring gear are also included in addition to an array of five carriers with different pin hole position errors. The transmission error results indicate that the phasing condition of the gear set is the most critical factor resulting in varying levels and numbers of modulation sidebands around the gear mesh orders. A novel method of measuring planet load sharing is developed to investigate the effects of mesh phasing and carrier pin hole position errors on individual planet loads. Strain gauges mounted directly on the planet pins are used to record the loads carried by the planets assembled in a fixed carrier continuously. The results of the tests indicate that the planet mesh phasing has no influence on the planet load sharing while the tangential pin hole position errors dictate the amount of load each planet carries. A single effective planet error parameter is shown to dictate the planet load sharing of four-planet gear sets. Finally, the orbital patterns of the planetary sun gear are recorded using a proximity probe measurement system designed for this study. These measurements show that a floating sun gear moves along a well-defined trajectory consisting of an integer number of trochoidal loops in an attempt to self-center the gear set. The exact number of these loops is shown to be dictated by the overall period of the gear set.
By analyzing the collected data, various conclusions are made at the end of the study in regards to the three system-level effects under investigation. It is shown that several of the test variables have design consequences and affect the behavior of the gear set. A list of recommendations for future work is also proposed to guide further investigations on this topic.