The fan is a mechanical device that functions to produce flow in a fluid, usually in the form of air. In fan industry, there is a drive system that is a series of gears that serves to reduce rotation. The gears often suffer damage, so the resulting rotation is not optimal. Thus, this study aims to measure the effectiveness of Time Synchronous Averaging (TSA) with Support Vector Machine (SVM) to examine damage to gears. This study uses the Support Vector Machine (SVM) method, to classify various variations of conditions from each gear fault. In this study using MATLAB 2018a software. There are 3 types of gear condition variations (normal, fault 1 and fault 2), each recording of 500 files per condition is performed. Grouped into two data, namely the original data and data with Time Synchronous Averaging (TSA). Each variation of conditions is extracted into a number of time domain statistical parameters and selected using the Relief Feature Selection. SVM classification is done by binary (two classes) and multi classes (many classes), using the Radial Basic Function (RBF) kernel function. The results of research conducted on several variations of this gear, namely the data without TSA treatment showed optimal classification with 100% accuracy. While the data that gets the treatment (after) TSA in the SVM classification shows results that are not optimal, namely with an accuracy rate of 90.9%