iSixSigma

Run Chart versus Control Chart

Six Sigma – iSixSigma Forums General Forums Tools & Templates Run Chart versus Control Chart

This topic contains 25 replies, has 14 voices, and was last updated by  Gurbani 9 years, 7 months ago.

Viewing 26 posts - 1 through 26 (of 26 total)
  • Author
    Posts
  • #53343

    Vargas
    Member

    I am hoping that someone could explain when to use control chart versus run chart. My thought is that if you have “enough” data, you should alwyas use control chart because it has more tests. Is that right?

    0
    #189687

    bbusa
    Participant

    Srini

    Run charts are ‘unstatistical’ – they have no statistical basis

    you cannot make conclusions of process behaviour from run charts and hence cannot make process adjustment . However , they are simple to construct and understand from a lay person’s perspective

    Just my opinion

    0
    #189710

    Joe B
    Participant

    Typically you start to monitor with a run chart and then judge if your process is stable and well distributed. If your data is well dispersed (use a histogram). then you can start to use a control chart limit to monitor. That tool and a stable process lets you predict the future and find process upsets.

    0
    #189712

    Andrew Banks
    Participant

    What you describe as a “run chart” I have learned as a “time series”.

    To me, a “run” chart is a time series with the median as the center line. There are statistical tests for whether the data shows too few or too many “runs” = above to below the median OR below to above. There is also a test for “clusters” = too many consecutive points on the same side of the median.

    On the other and, time series charts do not have any statistical basis – they simply track the metric on the y-axis versus a time-based x-axis.

    Another option I have seen used for new processes is a pre-control chart. It uses the target for the centerline and design specs, equipment limitations and “engineer’s intuiton” to lay out 1, 2 and 3 std dev “areas” in green, yellow & red on either side of center. As the first points are plotted they can be compared to the “ideal” and the Westinghouse rules can be loosely applied to detect the early performance of the process.

    Warmest Regards,

    0
    #189721

    Draper
    Participant

    Andrew Banks.

    Do you ever answer someones question correctly, or just ramble on with nonsense that is not pertinent?

    Don

    0
    #189724

    HopeOverExperience
    Participant

    Wrong. Run charts give you some statisitcal information to tell you in your data is stable e.g. clusters/mixtures or trends/oscillations. These value will tell you if your data is stable or not.

    0
    #189726

    bbusa
    Participant

    Clive T

    Wrong !

    Pl understand what ‘stable’ means statistically !

    Run charts can never tell you if the process is stable or not . There are no control limits on the run charts

    Pl ‘run’ to the nearest stats book

    0
    #189730

    HopeOverExperience
    Participant

    bbusa

    From my undertsanding your confusing “in control” with stable.

    Control charts will tell you if your data is in control while a run chart will tell you if your data is stable. Control charts do tend to give you better data but they do give you some statistical analysis of the data.

    0
    #189731

    Hody
    Member

    “Pre-control” charting, as one of its better descriptions was in Keki Bhote’s “World Class Quality” book ( a Shainin follower), is very different from SPC since it uses Spec limits rather than Control limits and fewer data points. It’s an early warning for process monitoring with fewer data points and is easier for the operators to implement…

    As I recall, once you have a capable process (Cpk>=1), you set the Green area in the middle 50% of the spec, Yellow bands are the upper and lower, but still in-spec. Reds are anywhere outside of the spec. Periodically check 2 parts: stop the line on any Red, call-in QC on 2 Yellows, else keep going…

    0
    #189732

    bbusa
    Participant

    May I please have a definition of a “stable” process ?

    0
    #189733

    Severino
    Participant

    bbusa wrote:

    May I please have a definition of a “stable” process ?

    Stable Process: a sequence of events in which livestock are corralled into an enclosure

    :huh:

    0
    #189734

    bbusa
    Participant

    Idiot

    0
    #189735

    qualitypupil
    Participant

    As I know stable and being in control are the same thing.

    0
    #189736

    bbusa
    Participant

    You are right !

    0
    #189741

    DMR
    Participant

    You are absolute right.

    RUN CHART gives a pictorial view of the process. Unless the specification limits are exceeded, you can never know how the process is performing.

    In CONTROL CHART, you can predict even before the process goes out control. In some situation, (types of data and hence type of control chart), within 1 sigma, it is possible to tell, if the process is out of control. This is not possible with RUN chart.

    So RUn chart is ok, if type of decision is not critical.

    0
    #189742

    Severino
    Participant

    bbusa wrote:

    Idiot

    So one bad joke merits a disparaging remark?

    0
    #189746

    Darth
    Participant

    Don’t worry too much about. Seems Stan has spawned a son. Along with the new format comes a loss of humor.

    0
    #189748

    bbusa
    Participant

    Humor – you un’Darth Stan(d)?

    0
    #189749

    bbusa
    Participant

    Humor – you un’Darth Stan(d)?

    0
    #189750

    Severino
    Participant

    Darth wrote:

    Don’t worry too much about. Seems Stan has spawned a son. Along with the new format comes a loss of humor.

    Darth –

    I appreciate your input. It would take far more eloquent use of the English language about my mental capacity for me to even blink an eye, let alone worry.

    Concerning Stan
    Beyond the usual forum drivel, there is a point of honor between you two that is unsettled. I myself have said things which were far more insensitive about people who were equally as undeserving so I do empathize with you, but I appreciate the value of a friend who defends you even if you yourself choose no defense. If you feel you’ve reached a point where reconciliation is no longer an option, then my only request is that you don’t throw your stones in a manner that suggests that I somehow harbor the same malcontent that you do as I do not share the same thoughts on Stan’s sense of humor.

    Again, I do appreciate your feedback so don’t interpret the second paragraph as anything more than a “leave me out of it.”

    0
    #189751

    Severino
    Participant

    bbusa wrote:

    Humor – you un’Darth Stan(d)?

    Nicely done.

    0
    #189757

    Darth
    Participant

    Now I’m confused. I think Stan has a sense of humor. I was reacting to BBUSA’s not finding humor in your cattle joke. Because of the revision to the website, Stan, Mike, HeeBee, Steveo and others have dropped off quite a bit. I had some time to kill so I thought I would try to tackle the new version. I imagine site volume has dropped a bit.

    0
    #189764

    B2B
    Participant

    Run charts are used to visualize your process over time without regard to specification limits or controls.

    Use a control chart to determine if your process in “in control” or within “specification” as defined by your customer/stakeholder. Control limits and spec limits are set to analyze process performance over time.

    0
    #189765

    bbusa
    Participant

    B2B

    Wrong !

    Spec. Limits are irrelevant to control charting – process stability is w.r.t control limits only :dry:

    0
    #189783

    Leader
    Participant

    The sad thing is that I have seen software (not Ms Mini, of course) that allows you to put spec limits on control charts. They’re trying to be helpful, like Microsoft….

    0
    #189816

    Gurbani
    Participant

    bbusa wrote:

    Run charts are ‘unstatistical’ – they have no statistical basis

    Wrong. They are a simple way of showing trends.

    Typically you start to monitor with a run chart and then judge if your process is stable

    Wrong. Run charts cannot tell if a process is stable.

    Another option I have seen used for new processes is a pre-control chart.

    Do not use these. They can be very misleading.

    Wrong. Run charts give you some statisitcal information to tell you in your data is stable

    Sorry, wrong again. Run charts say NOTHING about process stability.

    From my undertsanding your confusing “in control” with stable.

    Wrong again. A stable process is PREDICTABLE. This only occurs when a process is “in control”. Only a control chart can indicate this.

    It is obvious that there is a great lack of understanding of the basics of quality. Read Wheeler “Understanding SPC” and Advanced Topics in SPC”. Alternatively some basic training may help: http://www.q-skills.com

    0
Viewing 26 posts - 1 through 26 (of 26 total)

You must be logged in to reply to this topic.