The goal of this paper is to show that in order to achieve good localization performances with a modern Electronic Support Measure (ESM) system based on Time Difference Of Arrival (TDOA) and Frequency Difference Of Arrival (FDOA) measurements, it is necessary to have a precisely synchronized time base on every sensor. An application case is detailed to show how an imperfect synchronization causes degradations in the localization performances at the output of the system. It focuses on a simple scenario: two airborne receivers using TDOA and FDOA measurements to localize a stationary emitter. In the first section the synchronization process and its error model are defined. The second section presents simulations which show the influence of these errors on the estimation of TDOA and FDOA measurements. In the last section are discussed the final localization performances, based on the calculation of the Cramér- Rao Lower Bound (CRLB) using the previous corrupted estimates of TDOA and FDOA.