Several points: * Consider that your Cpk is probably different for both placing large and small components. * Probably your routine placement rate SB the speed you use to monitor the accuracy of your process. * Some people use a calibrated glass test board and glass components for this type of test. Fuji and others happily sell these items. * Before proceeding towards calculating Cpk, you need to demonstrate that your process is stable and in control. People commonly use control charts for this. * A sample of 30, maybe 25 is a good run size for your calculation of Cpk. * Consider that you are not plowing new ground and your Fuji rep may be a good source of information and tried techniques on your machines.
Finally for background, I reviewed a book on statistics [Messina "SPC for SMT �] in a SMTnet Newsletter within the past 6 months. It has a pretty good chapter on Cpk. A book focused on capability indices is S Kotz & CR Lovelace; "Process Capability Indices in Theory and Practice"; Edward Arnold; 1998; ISBN: 0340691778. The Kotz/Lovelace book is less approachable than the Messina book, but does provide at least as much as you'd ever want to know about capability indicies. A general net search will bury you with information on Cpk.
reply »