R
I too am foxed with processes with 5+ Cpk not being within specs.
I totally agree with what Spaceman Spiff says. Most processes are subject to a systematic drift due to tool wear. It pays to set the process not at the centre, but closer to one end of the spec. The systematic tool wear will then gradually take the process to the other end of the spec zone. That way you can take the best benefit by producing maximum quantity at one tool setup.
However you need to be careful to see that the process already has a Cp of 1.33+, so that wide random variations and outliers do not result in out of spec work. If Cp is lesser than 1.33, you will still have to set the proces at the spec centre and keep fine tuning the centering more often in the run.
I totally agree with what Spaceman Spiff says. Most processes are subject to a systematic drift due to tool wear. It pays to set the process not at the centre, but closer to one end of the spec. The systematic tool wear will then gradually take the process to the other end of the spec zone. That way you can take the best benefit by producing maximum quantity at one tool setup.
However you need to be careful to see that the process already has a Cp of 1.33+, so that wide random variations and outliers do not result in out of spec work. If Cp is lesser than 1.33, you will still have to set the proces at the spec centre and keep fine tuning the centering more often in the run.