Machine spindle dynamics and (axis of rotation) error motions may vary as a function of spindle speed due to gyroscopic effects, changes in bearing preload, centrifugal forces, and thermals effects. It is necessary to characterize these changes in order to fully define the spindle's performance. In this research, two different aspects of spindle performance are considered: a) spindle dynamics; and b) spindle error (SE) motions. The objective is to simultaneously measure the (potential) changes in both the error motions and dynamic response with spindle speed.This work is motivated by the influence of spindle performance on machining operations. Machining instability (chatter) leads to poor surface finish, high rejection rates, rapid tool wear, and, potentially, spindle damage. Stable machining conditions may be identified using well-known milling process models. To do so, the dynamics of the tool-holder-spindle-machine assembly as reflected at the tool tip is required. Here, the dynamics of an artifact-spindle-machine combination are measured at the tip of a standard artifact when the spindle is rotating. Tests are conducted at different spindle speeds to capture the speed-dependent changes in the spindle dynamics. Receptance coupling substructure analysis (RCSA) is then applied to predict the tool point response for arbitrary tool-holder combination in the same spindle. RCSA is used to first decouple the artifact dynamics from the measured artifact-spindle-machine assembly dynamics (to isolate the spindle contributions) and then analytically couple the dynamics of a modeled tool-holder to the spindle-machine in order to predict the tool point frequency response function (FRF). A speed-dependent milling stability lobe diagram, which graphically depicts the allowable axial depth of cut as a function of spindle speed, is obtained by identifying the changes in tool point dynamics with spindle speed.Spindle error motions, which describe the variable position and orientation of the spindle axis as a function of the rotation angle, can affect machined surface quality. Non-contact sensors (such as capacitance gages) may be used to measure the SE motions while the spindle is rotating. A multi-probe error separation method is used to accurately isolate the SE motions and the artifact form error. Tests are repeated at different spindle speeds to examine the associated effects. Together, the identification of the speed-dependent SE motions and tool point FRF will enable an improved capability to predict the milling performance for a given tool-holder-spindle-machine combination.In this research, the speed-dependent spindle dynamics and the SE motions for three different Haas TM1 machine spindles were studied. At a spindle speed of 3800 rpm, the critical stable axial depth of cut predicted using the stationary tool point FRFs was 6 mm while that predicted using the speed-dependent FRFs was 10 mm. Stable machining was observed at an axial depth of cut of 9 mm at a spindle speed of 3800 rpm. The results showed that incorporating the changing dynamics of the spindle in machining stability models improved the ability to predict chatter. Further, the dynamics and error motions of an NSK HES-500 high speed spindle were also measured.