Exposure to fugitive dust can contribute to several respiratory health problems, and proper sampling of fugitive dust is necessary to assess exposure. However, field sampling of soil dust encounters problems from spatial and temporal differences in soil properties, field operations, and meteorological conditions. To minimize these problems, we designed a dust generator that simulates dust generation from soil. The dust generator consisted of a rotating chamber where soil samples were loaded and tumbled, and a settling chamber, where airborne soil dust samples were collected. As standard operating conditions, we decided on 2 g soil mass, 10 min sampling time, and 20 rpm rotating speed, with a flow rate of 30 l/min, based on three common soil textures of loam, sandy loam and silt loam. To evaluate optimal operating conditions, we used mixtures of Joomoonjin silica sand and clay. Although the average $PM_{10}$ concentration of Joomoonjin silica sand was low, dust concentrations were increased by an increased content of clay. The dust concentrations were consistent across repeated experiments, and showed similar concentration profiles during the sampling time with mixtures of clay and sand (coefficient of variation was $13.6{pm}w;7.1%$). The results demonstrated that these standard operating conditions were suitable for the dust generator, which can be used to investigate variations in soil properties that affect dust production and potential potency of fugitive dust exposure.