# two sigma limits on control charts

Six Sigma – iSixSigma Forums Old Forums General two sigma limits on control charts

Viewing 27 posts - 1 through 27 (of 27 total)
• Author
Posts
• #151954

Darth
Participant

You react to more common cause signals than you would with 3 sigma limits.

0
#46110

Whitehurst
Participant

If two-sigma limits are used in place of the conventiional three sigma limits on the control chart, what does occur ?

0
#151955

Len
Participant

Control limits are always 3 sigma.  They are not probability limits … anything other than 3 has no meaning.

0
#151956

Whitehurst
Participant

What about alpha risk, beta risk or confidence level

0
#151958

Mikel
Member

Wrong

0
#151961

Darth
Participant
#151962

Len
Participant

Stan is correct.
Here is the definitive article.  It is essential reading.
http://www.spcpress.com/ink_pdfs/Wheeler%20Neave.pdf
I would love to see anyone who can come up with any intelligent criticism of this article, rather than the usual needling that occurs here.

0
#151964

Len
Participant

A good post by Rip, however clustering inside 2 sigma, while definitely indicating “something is wrong”, is unlikely to be a result of rational subgrouping … and Rip is a fellow who is smart enough to understand the correct use of this term rather than the stupid stuff taught in many six sigma course.  The reason for such clustering is more likely to be that pointed out by Deming … fear … leading to employees putting down false data.

0
#151972

Participant

You may find this link more useful.
http://deming.eng.clemson.edu/den/archive/98.02/msg00174.html

0
#151973

harry
Participant

If Stan is referring to action limits he is right: But many statisticians in the past have receommended the use of ‘warning’ limits set to 2 sigma. (Yes – and they calculated the values of the constants based on the distribution – not on the economic sample size.)

0
#152020

Len
Participant

Warning limits are a return to the bad old days of pre-control. Forget 2 sigma … and forget 6 sigma as well !!  Control limits are always 3 sigma.

0
#152031

stupidguy
Member

If control  limits are VOP and spec limits are VOC… why do we have 3 sigma for control charts and 6 Sigma for process capability?
Ideally our process should be in control as well as capable right? In that case why 3 here and 6 there?
Could be a stupid question..but plz answer.

0
#152033

Participant

When calculating the control limits, you calculate the control limits separately and so you add 3sigma on each side. Effectively the wide of the process is 6-sigma.
When you are calculating process capability, you are verifying how good is your process width compared to the specification width. Here you divide by 6 because you are comparing the overall process width, which is 3+3=6 sigma.
So there is no discrepency. Hope your doubt is cleared. Use may use the left side menu ‘Statistics & Analysis’ for more information on the same.

0
#152036

BB
Participant

Hi Harry ,
Could you please explain further : Why when n > 15  , we can forget the use of 3 sigma limits for sub-groups of the control chart. I want to learn.
Thanks ,
BB

0
#152034

harry
Participant

The warning limits I’m refered to have nothing to do with Shainin Pre-control, which does not use a Shewhart (Xbar and s) chart.
You can forget about 3 sigma limits for subgroups where n > 15 if you like – once you’ve obtained some experience actually improving industrial processes.
If you disagree with my last comment, put up an example of a process, and tell us what you did to reduce variation! This challenge applies equally to all the other aliases you use!!!!!

0
#152041

Allthingsidiot
Participant

Ajit
Well said.Please  continue  to  prove  that this  forum has several real  experts,and  only  pretenders like………

0
#152039

Allthingsidiot
Participant

More  Defects

0
#152040

Allthingsidiot
Participant

Your  name  will  be  “SATAN” if  you  add “a” between the  “S” and “T”.Does  this  related  to SS?

0
#152061

Darth
Participant

As my friend Stan would say, WRONG!!!!  Your response makes no sense at all.  Oranges are Orange and Cherries are Red.  Control limits measure process stability and Sigma Level measures process capability.  Sigma level looks at the relationship between the process performance and the closest spec limit.  The concept of Six Sigma being 6 infers that we want the closest spec to be at least 6 s.d. from the mean.  The control limit is intended to spot an unexpected departure from what a stable process would produce and has nothing to do with specs.  The 3 s.d. convention was empirically determined and has no basis in probability.

0
#152066

melvin
Participant

What objeective are you trying to accomplish through making this change?
Bob

0
#152069

BTDT
Participant

Joe:The rate of false alarms will increase by a factor of 11.5.Cheers, BTDT

0
#152078

Len
Participant

Darth,
Very good.  It is so unusual to see someone here who knows what he is talking about.  It is great to see someone who appreciates that control limits are not probability limits.
Six sigma is based on where you place specification limits.  A defect occurs when a variable goes outside spec.
The following are some of the six sigma nonsense: 3.4 dpmo, dpmo, opportunities for defects,  floating means,  “corrections to CL”,  “sigma levels”,  “3 sigma processes”, 99.7% points in control limits, short term, long term, Z shift, assumptions of normality, normal distibutions drawn on histograms.

0
#152079

Savage
Participant

You sound like another Wheeler fruit.

0
#152080

Participant

The question posed was:
“If control  limits are VOP and spec limits are VOC… why do we have 3 sigma for control charts and 6 Sigma for process capability?”
Darth:
For calculating the control limits we use Xbar +/- 3s.d, while for calculating the process capability we use (USL-LSL)/6s.d. Why do we use 3sigma in control limt and 6sigma in process capability?
“Sigma level looks at the relationship between the process performance and the closest spec limit.  The concept of Six Sigma being 6 infers that we want the closest spec to be at least 6 s.d. from the mean”.
Probably you misread the word ‘process capability(Cp)’ as ‘process capability index (Cpk)’. For process capability (Cp) calculation, we never know how far the process is from the spec. limit, we only know, how good the process width fits into the specification width. So there might be a situation in which the process capability is very good and the points are not within the spec. limits. This will reflect in the process capability index (Cpk).
I agree to your reasoning that there is no relation in origination of the formula between the 3s.d used in control limits and 6s.d. used in Cp calculation. At the same time, Cp has nothing to do with the specification limits. You can change the spec. limits but maintain the same spec. width and it willnot get reflected in the Cp.
So the reasoning given by you doesnot apply to the question posed.

0
#152084

accrington
Participant

What is a Wheeler fruit? If you disagree with what the previous poster said, or with Donald Wheeler, please give us the benefit of your superior knowledge, and tell us where we are going wrong. If you want to be taken seriously, use debate.
If you just want to throw insults, exchange posts with our concrete – headed chum, I’m sure he’d love it!

0
#152088

Ashman
Member

Insults are all that small minds like Matts are capable of.
Concrete thinking like that of Darth and Len is hard to beat.

0
#152114

Darth
Participant

Ajit, my apology on my misread of the post.

0
Viewing 27 posts - 1 through 27 (of 27 total)

The forum ‘General’ is closed to new topics and replies.