Specs vs Control Limits Confusion

Six Sigma – iSixSigma Forums Forum Basics Welcome Specs vs Control Limits Confusion

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
  • #53620


    Please help me with my (and my teammate’s) confusion.

    We are in the beginning phase of establing control chart for a certain tool parameter in our process. Initially, before we started this improvement, we were only following specs limit as our basis of process being within “control”.

    But now that we already have computed & applied control limits to our charts, other teams are reacting: do we eliminate or specs limits now? do we change our checksheets from Specs to Control limits values?

    I am thinking, we have targets that should be our basis of calibrating or setting the tool, not the specs (like if the parameter is higher that the LSL, its enough for them. They aim on the specs limits, not the target.) But how do I explain this to them more technically & with a stronger foundation?

    I need your help badly.




    Hi Anne,

    In a word they are “Different”. Like apples and oranges. Control limits are used to assess whether a process, or a process variable is operating in control +/- 3 sigma limits, X bar, R chart, etc.

    Spec limits are used among other tings to determine if a product part or process is meeting (in most cases) a customer generated specification, although it can just as easily be applied to an in house specification. Such as the O.D. of a cam shaft 50mm +/- 8mm.

    More detailed info can be found by searching this discussion forum or as a minimum using the dictionary of terms.

    Hope this helps a little….




    There is no relationship between control limits and specification limits. None. The data will deterimine the CL of the process, the customer / business will determine your SL or target values. Use control charting in conjunction with capability testing to determine where your process is relative to capability and stability. Your process can be in any one of these four states (ideally – fully capable and stable) and understanding this is critical in determining what you do next.

    If your process is devoid of special cause (aka stable) then you can move forward analytically, looking at process capability and using other inferential statistical methods to optimize the process Y and reduce defective output. In addition, CC use is often continued to show “before and after” effects on the project Y to show the effectiveness of your change on the central tendency / variability in the process.

    I would highly reccomend you brush up on your technical understanding asap before you attempt further explanation.

Viewing 3 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic.