Downloaded from https://www.everyspec.com MIL-HDEK-68S(AT) - TW II-4. ~ ~ * - Continued. Pitfall Possible Commqwncea Not selecting subgroups on a real-time basis. Assignable causes become ~ of within subgroup variation caueing unnece8f3arilywide control llmit6. This results in failure to detect amignable causea when they are ~sent. Subgroups are too close together during Process Capability Study. Inability to evaluate the stability of the process. Subgroups are too far apart, too much time between subgroups. Shifts may occur and not be detected and large nuuiberof defective parts produced. Operatom are aware of which samples will be used for the subgroup.
will not know true process averages and variation. The process may appar in-control when it is actually not. Using subgroups when4individuals are appropriate for a proceea with a low production rate. a. Assignable causes become part of within subgroup variation causing unnecessarily wide control limits. This results in failure to detect assignable causes when they - present. b. Shifts may occur and not be detected and large nuniberof defective pts produced. The average of the daily percent defective is used for the process average (~) rather than the formula: 5= More weight will be given to days with lower production rates. Those might be daya with highest percent defective.
ts Total Inspected No control plotted on limits or center line charts. No indication when to take action. Action taken may be based upon personal choice. Plotting specification limits on a control chart with normal control limits. tinfusion aB to whether action should be taken. Process may appear within specification when it is not. 44 - . ---- - . - . . . . . ---