This combined experimental and computational study is an extension of prior efforts by the present authors to provide a physical explanation of the ability of a particle detection technique based on the acoustic signature of airborne particles. As a particle-laden flow accelerates through a converging-diverging tube consisting of a capillary (called the acoustic transducer), the suspended particles cannot follow the rapidly changing flow velocity. Consequently, vortices are generated in the wake of the particle, which excite the air column inside the transducer tube producing noise. Experimental results show how the frequency content of the acoustic signatures relates to the fundamental frequency of the transducer's air column. The acoustic transducer is able to detect micron-sized particles and the sound intensity is a function of flow rate but not of particle size. The ability of the transducer to determine particle concentration (as low as a few ppl) is also experimentally shown and compared to data obtained from a commercially available aerodynamic particle sizer. A computational model is developed which incorporates detailed treatment of air compressibility as well as the variability in Reynolds number and particle drag coefficient. Computational results help determine the location of the acoustic disturbance, which is at the entrance of the capillary section of the transducer where the flow speed is nearly sonic and the acceleration is the greatest throughout the flow. (C) 2008 Elsevier Ltd. All rights reserved.