CNC Benchtop Lathe

My most complex design project to date, this CNC benchtop lathe is designed to turn aluminum and steel with at least 50um precision. Lead screw actuated x-feed and z-feed mechanisms can be controlled with dials or via stepper motors for CNC operation. The z-direction translation system comprises an aluminum carriage, bronze bushings, and steel rails. The carriage mounts to one of the guide rails via a 2-DOF flexure to relieve overconstraint. A 4-DOF flexure attaches the lead screw drive nut to the carriage, again to eliminate overconstraint. The x-direction translation system (cross feed) is a 1-DOF flexure bearing stage, which mounts the tool post. The spindle contains a steel shaft, two tapered roller bearings (for better shaft moment resistance, and to avoid thermal runaway), an aluminum housing, a chuck, and is driven with a V-belt transmission.

The design of the machine was formulated around a homogenous tranform matrix (HTM) model to track alignment/machinining errors and component deformations under turning forces, ensuring that the required precision would be met. I personally designed the three flexure mechanisms using finite element analysis (SolidWorks and Abaqus), contributed to the design of the spindle, x-feed and z-feed drive mechanisms, and contributed to the development of the HTM model. Nearly all parts were machined at MIT on the mill, lathe, and water jet, after which the lathe was assembled and tested.

Runout and eccentricity measurements of the spindle were taken to the optimize spindle bearing axial preload. CMM (coordinate measuring machine) measurements of the cross feed flexure stage were taken while moving the x and z axes to verify that x, y, z, pitch, roll, and yaw error motions of the tool were acceptable and shimming was not required. Turning tests showed a taper error of 17um, an eccentricity and runout error of 16um, and mean diametric error of 26um, subsequently meeting the precision requirement with 35um RMS error. The lathe also successfully passed three death tests. First, the chuck was hit repeatedly with a sledgehammer to verify proper spindle bearing preload. Second, a person stood on the flexure stage and was manually translated in the along the x and z axis to check that flexures were sufficiently stiff in their degrees of constraint. Finally, the lathe was dropped from 1m to check that the flexures contained appropriate hardstops for avoiding buckling and yielding under the impact.

Website created with Sandvox. Hosted on Amazon S3.