Rotating machinery is widely used in industrial applications and it is important to ensure its smooth running in order to prevent system failure. Therefore, gear health monitoring becomes vital and gears can fail due to various faults like shaft misalignment, crack, missing tooth, wear etc. In this paper, condition monitoring of a spur gear assembly is done with normal gear and also with faulty gear with cracked tooth. The dynamic simulation of crack propagation at the root level of a single tooth of the spur gear is done with Abaqus finite element analysis software using the contour integral method. The gear transmission is subjected to different input torque and also run at different speeds and loads. Vibration amplitude obtained from the simulation in terms of velocity is used for fault diagnosis. The velocity data captured is passed to the Symmetrized Dot Pattern (SDP) algorithm developed in MATLAB that generates visual polar plot to identify the different gear faults easily. This plot is a unique signature for the given gear health condition. Using the developed gear fault diagnosis technique, we are able to successfully identify the health of the gear assembly, including the faulty condition i.e. crack in the gear. The computational time required to run this algorithm is also less compared with the other fault diagnosis techniques. © 2019 IOP Publishing Ltd. All rights reserved.