D
We do CNC lathe/mill work. I know that one of the best control methods for this type of work is SPC. However, I would like to get your input on setting up sequential sampling frequencies. My definition of sequential sampling is: Operator inspects 1 out of every X pieces. (Is that the correct definition of sequential sampling?)
Our operators inspect parts by this methodology. The frequencies we've setup are determined today by evaluating the tolerance and what we think would be a good frequency would be to prevent too much scrap. I feel this method isn't the correct way to do things. I would like to move in a direction where these frequencies are more statistically driven. From what I've seen, inputs to deterermining this would be cpk and PFMEA; are there any other tools/methods to determine this? Any advice on how to use these tools to develop these frequencies with these tools?
From a documentation standpoint what should be recorded? I would think that the frequency we define on our control plans dictates how many times we should record dimensions. I saw this thread on sampling evidence requirements: Sampling Evidence Requirements - Pursuing AS9100. I was asked by management if we could provide multiple frequencies on our control plans (One frequency for how often operator should check the dimension and another frequency for how often to record the dimension). I feel that not recording dimensions every time the operator checks and just at specified document dimension frequency would not properly control the process. Any thoughts about this? Is it feasible to specify multiple frequencies as said above?
Your help is greatly appreciated.
Our operators inspect parts by this methodology. The frequencies we've setup are determined today by evaluating the tolerance and what we think would be a good frequency would be to prevent too much scrap. I feel this method isn't the correct way to do things. I would like to move in a direction where these frequencies are more statistically driven. From what I've seen, inputs to deterermining this would be cpk and PFMEA; are there any other tools/methods to determine this? Any advice on how to use these tools to develop these frequencies with these tools?
From a documentation standpoint what should be recorded? I would think that the frequency we define on our control plans dictates how many times we should record dimensions. I saw this thread on sampling evidence requirements: Sampling Evidence Requirements - Pursuing AS9100. I was asked by management if we could provide multiple frequencies on our control plans (One frequency for how often operator should check the dimension and another frequency for how often to record the dimension). I feel that not recording dimensions every time the operator checks and just at specified document dimension frequency would not properly control the process. Any thoughts about this? Is it feasible to specify multiple frequencies as said above?
Your help is greatly appreciated.