Sand is extensively used in geotechnical engineering applications due to its abundance in the field. In general, sands are subjected to very low strain rates in most geotechnical engineering applications; however, during air or surface blasts, sands are subjected to high strain rates and their response needs to be studied in detail. The recent years have seen an increase in the usage of split Hopkinson pressure bar (SHPB) device to study the response of sand at very high stresses (on the order of MPa) and strain rates (10~1000/s). Researchers have shown that the particle size in sand influences its mechanical behavior at low strain rates. However, the effect of particle size on the stress-strain response of sand obtained from an SHPB device where particles are subjected to high strain rates is seldom studied. The particle size could also have an influence on the stress-equilibrium condition which needs to be satisfied by the specimen for the stress-strain analysis from the SHPB to be valid. However, the afore mentioned criteria are seldom studied and need to be studied in detail. In the current study, the SHPB device is modelled using a commercial software package LS-DYNA with a coupled finite elements-discrete element method. The sand specimens are modelled using polydisperse (spheres with varying sizes) and monodisperse discrete spheres. These simulations are used to understand the variation of stress-strain response of sand as a function of particle sizes. The study also provides insight into the effect of friction at the boundaries on the stress-strain response of the specimen. The results indicate that the specimen having larger polydisperse spheres provided similar stress-strain response as the specimen having smaller polydisperse spheres. However, the monodisperse specimen provided a softer response compared to the polydisperse specimens.