Extensive experimental data was collected for CBN surface grinding of M2 tool steel to determine the relative grinding performance of three different vitrified CBN abrasive grit sizes. The results define the relationships between the grinding forces, the material removal rate and the resulting specific energy, while providing an evaluation of the ground surface characteristics including surface finish, microstructure, hardness and residual stress. The interaction of grinding process inputs including wheel grit size, workpiece velocity and depth of cut are considered, and a series of single factor tests and a 23 factorial test are conducted. The grinding forces increase linearly with increasing material removal rate for the range of parameters tested. As expected, decreasing abrasive grit size increases the grinding forces and required power while the surface finish is improved; however, unlike conventional abrasive grinding, the tangential grinding force and resulting power and specific energy differ only slightly. The results of the factorial test indicate that abrasive selection, workpiece velocity and actual depth of cut, along with second order interactions between abrasive selection/workpiece velocity and depth of cut/workpiece velocity significantly influence the normal force response, while the tangential force response is influenced by the main effects of workpiece velocity and depth of cut and slightly by the abrasive selection. CBN grinding leads to compressive residual stresses at the workpiece surface, while grinding at similar conditions with conventional abrasives leads to tensile residual stresses. For high material removal rate grinding where visible oxide layers appear on the workpiece surface, the evaluation of near ground surface metallurgy indicates formation of untempered martensite at the surface and a transition region of overtempered martensite between the workpiece surface and the bulk material. It is shown that this burn occurs at the same effective grinding temperature for each abrasive. The talk concludes with comments about applying this burn limit model to interpret observations of workpiece burn in OD grinding.