Abstract
Milling experiments have been performed to investigate the dependence of surface finish s on cutting speed v feed f; and depth of cut d|Regression analyses were carried out for a nonlinear model and for a linear model obtained therefrom by logarithmic transformation. The good agreement between the two sets of solutions confirms the adequacy of the model and the linear solution values are therefore commended as suitable 'initial guesses’ for an iterative solution of the nonlinear case. In addition, small deviations of the 'controlled’ variables from values suitable for coding for a three-level factorial design were found not to be important. Finally, the dependence of surface finish on feed rate f was in fact so weak that the suppression of this parameter had little effect on the exponents of the other parameters. Overall the relationship could therefore be well represented by sα(v/d)15/4.