Dynamic compressors operating region is mainly constrained by fluid-dynamic instabilities occurring at low mass flowrate conditions, such as surge and rotating stall. This work presents a vibro-acoustic experimental investigation on a centrifugal compressor of an automotive turbocharger aimed to identify and confirm some surge precursor values in correspondence of its inception conditions. The experimental campaign was carried out at the University of Genoa and developed on a vaneless diffuser turbocharger exploited for the pressurization of an innovative solid oxide fuel cell (SOFC) emulator. The investigated turbocharger is coupled with a pressure vessel for a former emulation activity on a pressurized SOFC. In this kind of plants, the joint effect of large volume size downstream of the compressor makes more complex the dynamic behavior of the whole system during transients, thus significantly increasing surge onset risk. The activity or the main goal is to obtain a suitable quantitative indicator capable to detect in advance surge inception by relying only on vibrational and acoustic system response. Several transient operations starting from a compressor stable condition to surge instability onset were performed at different initial rotating speeds by progressively closing specific valves in the air line. When moving close to the surge line, vibro acoustic signals were acquired at a high sampling rate to detect variations in compressor blade passage phenomena due to possible interactions with rotating stall inception. Meanwhile, the trend of pressures, temperatures, and mass flow rates measured in specific plant sections were acquired at a lower sampling rate to obtain a link between the compressor vibro-acoustic and performance behavior. Cyclostationary analysis and several postprocessing methods in time, angle, and frequency domains were performed on microphone and accelerometer acquired signals to provide innovative diagnostic and predictive solutions (precursors) able to warn the incoming of surge compressor instability with cheap and not intrusive sensors like microphones and accelerometers.