Hi All,

I have just looked into my workplaces SPC System:whip:, And I have wrapped:agree1: my head around of how it works the only problem im having is i don't know if it is a six sigma system or three sigma. How can you tell the difference? :o:thanx:

Kind Regards,

Omeil

Process charts are generally set up in six bands representing all the areas of probability, your upper control limits and your lower control limits.

Six, versus three sigma is describing where most of your data points lie. Are they spread out all over the place, or are they centralized?

This site (http://en.wikipedia.org/wiki/Six_Sigma) has some more explanation.

What I have is control charts representing 7 lines-

-USL----------------------------------

---------------------------------------

---------------------------------------
----------------Target----------------
---------------------------------------

---------------------------------------

-LSL-----------------------------------

each deviation is around .40, then from the moving average and the daily average i calculate the CPK.

Generally at the start i thought it was three sigma due to it having 3 standard deviations on each end.

Thanks, Omeil

Generally at the start i thought it was three sigma due to it having 3 standard deviations on each end.

Thanks, Omeil

Like Jennifer said, it all depends on if you are looking at the entire distribution, or one side of the distribution. One one side, three sigma will cover 99% on one side, and 3 sigma on the other side.

As to your information you provided, it looked broken up a bit. You might consider providing it as a JPG or something.:)

I have attached an ex. had to remove some text due to confidentiality purposes but the SPC data is there.

Thx again :)

First, you need a process that is stable and in a state of statistical control. This process is not in control, so any estimation of the process capability will not be valid for the next sampling period.

Second, your report section has a metric called Process Sigma = 0.438. This is the sigma level of your process. You should read up on the calculation used to determine this metric for your software. Typically this number should be 3*Cpk.

A 3-sigma process uses 100% of the tolerance. A 6-sigma process uses 50% of the tolerance.

Oh im aware of the control issues in this area of SPC, and its a work in progress :). CPK is increasing by the day, all through training. So if the process is in control the cpk/3 would equal something very close to the sigma? from the sample provided am i following six sigma or three or does it depend on the calculation of the cpk?

thanks,

omeil

I guess it would not hurt at this point to ask what the process is, and if there are any "events" - such as adjustments - that occurred during the data period? :cool:

All i can say is its food manufacturing. from the data you can see there is the person was making to large adjustments to the process not realising it is effecting quaility, now i have put an automated cpk calculation in front of them, so now they are actually trying to work better because they are using the cpk as a game score for there own motivation i guess. Lately the process has become heaps more stable cpk is usually over above 1 which is where it is supposed to be for this process. :cool:

Oh im aware of the control issues in this area of SPC, and its a work in progress :). CPK is increasing by the day, all through training. So if the process is in control the cpk/3 would equal something very close to the sigma? from the sample provided am i following six sigma or three or does it depend on the calculation of the cpk?
I got in too big of a hurry earlier. The correct formula is 3*CP or 3*Cpk.

A normally distributed process that uses 100% of the tolerance has a mean on nominal, +/- 3 StdDev = the Upper and Lower specification limits and a Cp/Cpk=1. The sigma level for this process would be 3*1 = 3 sigma. A process with a Cp/Cpk = 2 would be a 3*2 = 6 sigma process.

Note: This does not consider the controversial +/- 1.5 standard deviation shift required for the 3.4 PPM that is discussed by the proponents and opponents of six sigma. The importance of that has been overblown for years and is not really relevant to the six sigma philosophy. You can use it or ignore it at your discretion.

All i can say is its food manufacturing. from the data you can see there is the person was making to large adjustments to the process not realising it is effecting quaility, now i have put an automated cpk calculation in front of them, so now they are actually trying to work better because they are using the cpk as a game score for there own motivation i guess. Lately the process has become heaps more stable cpk is usually over above 1 which is where it is supposed to be for this process. :cool:

Sounds like you are on the right track! For most food manufacturing processes, reducing variation is a good thing - but it goes back to the incoming materials, too. If the incoming materials are not consistent, your attempt for process control may not yield consistent or acceptable output. It can be tricky - it's that organic thing. But, still, using SPC as feedback to the employee (a key part of the process) is a good start! Congrats! :agree1:

All i can say is its food manufacturing. from the data you can see there is the person was making to large adjustments to the process not realising it is effecting quaility, now i have put an automated cpk calculation in front of them, so now they are actually trying to work better because they are using the cpk as a game score for there own motivation i guess. Lately the process has become heaps more stable cpk is usually over above 1 which is where it is supposed to be for this process. :cool:
My first success in applying SPC and improving capability was simply getting the operators to stop making unnecessary adjustments. They were reacting to every single deviation and increasing variation just like the famous bead experiments.