Creating a c-chart spreadsheet to use in my department

R

riot_

Hi! My name is Manuel:

i'm working in create a c-chart spreadsheet to use in my department, but i've some doubts about this and hope you can guide me a lil': "what I'll meassure or control are ""Slubs"" and for this I know that is better use a chart for attributes, but the last thief in the department use to use a X-R chart for this.

this is for a textile company and we are thinking get our ISO/TS certification, hope someone can share a spreadsheet for this.
 

Stijloor

Leader
Super Moderator
Hi! My name is Manuel:

i'm working in create a c-chart spreadsheet to use in my department, but i've some doubts about this and hope you can guide me a lil': "what I'll meassure or control are ""Slubs"" and for this I know that is better use a chart for attributes, but the last thief in the department use to use a X-R chart for this.

this is for a textile company and we are thinking get our ISO/TS certification, hope someone can share a spreadsheet for this.

Just as an "aside" comment. ;)

Please note that ISO/TS 16949:2009 states the following:

7.1.2 Acceptance criteria
Acceptance criteria shall be defined by the organization and, where required, approved by the customer.
For attribute data sampling, the acceptance level shall be zero defects (see 8.2.3.1).

Emphasis mine.
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
The c-chart is strictly used for counting events that are expected to be independent from each other. The c-chart is based upon the Poisson distribution, and in queuing theory is used for a "random arrival process".

The big difference mathematically with the c-chart is that you use the standard deviation of the average as the estimator for the standard deviation.

Examples of things likely to be "poisson" include counting the following per some fixed unit of time:
Defects
Injuries (where only one person in injured)
Events
Arrivals of customers at a bank
Number of cars passing a certain location

Anything that involves a count per varying unit (injuries per 200,000 hours) or a measurement (dollar value of events per month) would not be a c-chart.

Bottom line is
1. Do you have a good theory that what you are measuring ought to be Poisson
2. Does the measured standard deviation come close to the square root of the average.
 
R

riot_

Thanks for your explanation, Stijloor, with that note of the 7.1.2 I can give a better response the next time.
 
R

riot_

hi Steve, I understan what you are saying, and is logical, but I've got a little confusion, this due the following criterial of xbar-r chart:

1.- The sample size is relatively small (say, n ≤ 10— and s charts are typically used for larger sample sizes) [slubs are very small defect].
2.- The sample size is constant [the defect is often in the bobbins, just the size is what matters].
3.- Humans must perform the calculations for the chart [indeed, my operators are trained for that].

but also, you are saying that the next:

Anything that involves a count per varying unit (injuries per 200,000 hours) or a measurement (dollar value of events per month) would not be a c-chart.

what I'm trying to measure are "slubs" in the bobbins to determinate the quality and procesability of the yarn, and really I don't know which could be better option (c-chart against xbar-r chart) to know that.

hope you can get my doubt.

PD. do you have any c-chart spread sheet!?
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
hi Steve, I understan what you are saying, and is logical, but I've got a little confusion, this due the following criterial of xbar-r chart:

1.- The sample size is relatively small (say, n ≤ 10— and s charts are typically used for larger sample sizes) [slubs are very small defect].
2.- The sample size is constant [the defect is often in the bobbins, just the size is what matters].
3.- Humans must perform the calculations for the chart [indeed, my operators are trained for that].
PD. do you have any c-chart spread sheet!?

statement 1 has no effect on use of c-chart
statement 2 supports the use of the c-chart
statement 3 - the c-chart standard deviation is one of the simplest to calculate - it is simply the square root of the average, and the control limits, once calculated, are extended forward as a horizontal line.

For c-chart spread sheet, see http://www.efcog.org/wg/esh_es/Statistical_Process_Control/SPC_Trending_Primer/daytwo/11%20C%20Chart%20Day%20Two.pdf or attached.
 

Attachments

  • C Chart Demo 12152010.xls
    35.5 KB · Views: 239
Top Bottom