With a view to enabling additive manufacturing (AM) processes, today, open-source, low-cost 3D printers are systems with great potential. However there is a significant lack of scientific data on the performance of open-source 3D systems and on the selection of adequate process parameters that can help to improve the quality of the parts. The purpose of this paper is to assess the effects of the main process parameters on the dimensional accuracy of a specific open-source 3D printer, the RepRap Prusa-Mendel I2. This study consisted of a benchmarking part, involving elementary shapes representing a series of different features. By means of a full factorial DoE (Design of Experiments), with three factors (layer thickness, deposition speed, and flow rate), three levels, and three replications, 81 parts were obtained. Subsequently, a laser scanner (D700 Laser Scanner—3Shape, Denmark) was used as high resolution reverse engineering system in order to evaluate the variation between real parts and nominal geometry. The impact of the main process parameters was evaluated and optimal combinations were analyzed. On the basis of the results obtained in the experiments, practical suggestions for the settings of common process parameters were formulated. Test results serve to improve the quality of AM parts through the most appropriate selection of the main process parameters.