Control Limits
Six Sigma – iSixSigma › Forums › Old Forums › Software/IT › Control Limits
- This topic has 9 replies, 10 voices, and was last updated 17 years, 2 months ago by
Harini V.
-
AuthorPosts
-
March 18, 2005 at 12:39 pm #26274
Bob LoweParticipant@Bob-LoweInclude @Bob-Lowe in your post and this person will
be notified via email.I am new to SS and am using minitab to run an IM-R chart. I read that the control limits are set at +- 3 sigma from the mean. Yet, when I calculate this from the descriptive statistics that minitab gives me, it is never exact. It sometimes gives me 2 sigma, sometimes 3 sigma plus n and other times no logical explanations….Can someone please explain how minitab calc’s the limits?
0March 29, 2005 at 6:53 pm #63847
francobaezParticipant@francobaezInclude @francobaez in your post and this person will
be notified via email.read the help files from minitab, it is explained
0March 29, 2005 at 8:02 pm #63848I did check the help files in minitab and it references that the limits are set at 3 sigma. Reality is that the limits are set at 2.66 times mean of the moving range plus mean of the individual value chart. Do not understand this logic.
0March 29, 2005 at 8:23 pm #63849Stat > Control Charts > Xbar, Individuals, Xbar-R, Xbar-S, I-MR-R/S (Between/Within), I-MR > Tests
Stat > Quality Tools > Capability Sixpack (Normal), Capability Sixpack (Between/Within) > Tests
Lets you select a subset of the eight tests for special causes shown in Eight Tests for Special Causes. Each test detects a specific pattern in the data plotted on the chart. The occurrence of a pattern suggests a special cause for the variation, one that should be investigated.
When a point fails a test, Minitab marks it with the test number on the plot. If a point fails more than one test, Minitab marks it by the lowest numbered test, and prints a summary table with complete information in the Session window. See [5] and [25] for guidance on using these tests that detect specific patterns in the data plotted on the chart.You can adjust the sensitivity of the tests with Stat > Control Charts > Define Tests.
Subgroup sizes must be equal to perform the tests.
Stat > Control Charts > Define Tests
You can define the sensitivity of the tests for special causes used with quality control charts. The test definitions stay in effect until you restart Minitab. The range of acceptable values you can enter depends on the test number, as shown below.
Test K can be (default in parentheses)1 One point more than K sigmas from the center line 1-6 (3)2 K points in a row on same side of center line 7-11 (9)3 K points in a row, all increasing or all decreasing 5-8 (6)4 K points in a row, alternating up and down 12-14 (14)5 K out of K + 1 points in a row more than 2 sigmas from the center line (same side) 2-4 (2)6 K out of K + 1 points in a row more than 1 sigma from the center line (same side) 3-6 (4)7 K points in a row within 1 sigma of the center line (either side) 12-15 (15)8 K points in a row more than 1 sigma from the center line (either side) 6-10 (8)
0March 30, 2005 at 6:02 am #63850
ChaturvediParticipant@PankajInclude @Pankaj in your post and this person will
be notified via email.The value of UCL and LCL are computed as X bar + 3 sigma and X bar – 3 sigma. Here the sigma is equal to the standard deviation of the sample / root n. To know more on this read book on quality control. This is the basic of control charts.
0March 30, 2005 at 7:00 am #63851Control limits are calculated depending on the type of control chart you are using. In a typical X-bar R chart the constant A2 is used to calculate the limits. For the I-MR chart of individuals, they use a constant called E2 because there is only one sample in each subgroup. For individuals, use the constant E2=2.66 multiplied my R-bar, or the average of the moving range chart. This +/- the mean will get your limits.
For all the E2 constants, you should be able to get on the internet. They are also in the BB Pocket Guide and most books. Or, see this:
http://www.qualityadvisor.com/sqc/formulas/x-mr-f.htm
0March 30, 2005 at 2:57 pm #63855
Paul AaronParticipant@Paul-AaronInclude @Paul-Aaron in your post and this person will
be notified via email.Give Minitab a call at 814-231-2682. They are very helpful.
0April 1, 2005 at 4:53 am #63858
Dale DoyleParticipant@Dale-DoyleInclude @Dale-Doyle in your post and this person will
be notified via email.Matt’s answer is correct in that an IM-R chart uses a constant in the formula. Keep in mind control limits are based on the data. Because the IM-R is looking at one data point to the next, we loose much of the power of the “C L T”. It will take a number of data points for the contol limits to level out. Mini Tab calculates the CL each time a new data point is entered. Depending on the type of data you are collecting you may want to check to see if this chart is the best fit for the process you are charting.
0April 1, 2005 at 1:24 pm #63859First of all you have to understand the logic of the control charts. By means of control charts you can compare long term variation against short term variations. In the case of X-MR chart, the short term variation is estimated by a difference between two or more successive individual values moving with time. If the long term variation is significantly larger than short term variation (existence of variation other than common casue) then the dfference between the ways of control limit calculation will be bigger.
0April 4, 2005 at 4:48 am #63862
Harini VParticipant@Harini-VInclude @Harini-V in your post and this person will
be notified via email.Please refer to the book by Donald Wheeler “Understanding Variation – The key to manageing Chaos”.
For IMR charts, with adjacent values, n=2, the k value becomes 2.66, which is nothing but 3/Hartley’s constant. Replacing Harley’s constant (1.128) for n=2 will give 2.660 -
AuthorPosts
The forum ‘Software/IT’ is closed to new topics and replies.