What are your Cp and Cpk Values?
Six Sigma – iSixSigma › Forums › Old Forums › General › What are your Cp and Cpk Values?
 This topic has 17 replies, 17 voices, and was last updated 15 years, 6 months ago by mand.

AuthorPosts

May 6, 2001 at 4:00 am #27256
I am doing research into Cp and Cpk and was wondering if some of you could tell me the Cp and Cpk values for your products?
0May 6, 2001 at 4:00 am #66532
Ken MyersParticipant@KenMyers Include @KenMyers in your post and this person will
be notified via email.Tom,
Don’t you mean process(es)?
0May 7, 2001 at 4:00 am #66542
Grant BlairParticipant@GrantBlair Include @GrantBlair in your post and this person will
be notified via email.I retired from Monsanto’s plant in Greenwood 4 years ago, but when I worked there in the 1980’s1990’s, we had a plantwide goal of a Cpk>3.0 for all processes.
Doesn’t mean we always achieved it, but the effort was
worthwhile.
The basis for this was not Six Sigma theory, but Chebysheff’s inequality…our plant manager at that time was a Deming advocate, a former corporate quality director, and a darn good statistician.0May 7, 2001 at 4:00 am #66543
Andy BrodyParticipant@AndyBrody Include @AndyBrody in your post and this person will
be notified via email.In the paint industry there is very little statistical opportunities for improvement (processwise). At least in our company. We are 149 years old, a custom manufacturer of coatings for OEM manufacturers. We have several thousand formulas and it is rare that we make the same formula once in a month, so a statistically valid study is not really possible. The best we were able to is to measure the output results of what is called a sandmill which breaks down pigment particles to the customer specifications and is measured by maximum acceptable size on a gauge which is highly subjective in reading with measurement increments that exceed its ability to accurately read the maximum particle size coming oput of the mill. However, this mill and gauge (called a grind gauge)is an industry standard used by manufacturers and customers alike. This process combination has proven to be amazingly successful.
Our cpk was calculated by running one batch. After the customer spec. was attained, a grind reading was taken every two minutes until the batch was completed (100 readings were taken). The cpk was 2.02. What is deceptive is that once the acceptable grind is obtained,it very rarely changes unless an outside force acts upon it.
0May 7, 2001 at 4:00 am #66545
Larry SeibelParticipant@LarrySeibel Include @LarrySeibel in your post and this person will
be notified via email.The general rule of thumb states that if a Cpk value of a process is less than 1.33, then the process is incapable of producing a repeatable part. Most of our injection molding processes yield a Cpk value of around 1.0. That means that I, as a black belt, have a great deal of work to do to make the processes predictable, then I can proceed with DOE’s and find the optimal preocess settings. Also, remember this, there is a difference between Cpk and long term process capability. That difference is long term is typically a 1.5 sigma shift for the worse, accounting for white noise (common cause variation. Cpk is basically a snap shot of what you have going on right now, unless of course you have the time to take data ofer a long period of time and are considering ALL variation (Common and Special Cause). Hope this helps!
0May 7, 2001 at 4:00 am #66547It doesn’t sound like that Cpk is valid at all. If I understand your posting correctly, you have a completely unreliabel gauge. Have you done any measurement system analysis? Saying it is industry standard is a cop out if you can’t get accurate and reliable readings.
0May 7, 2001 at 4:00 am #66549
AnonymousParticipant@Anonymous Include @Anonymous in your post and this person will
be notified via email.A industry wih little opportunity for process improvement? Sounds like a company in jeopardy!
0May 10, 2001 at 4:00 am #66584
Joe PeritoParticipant@JoePerito Include @JoePerito in your post and this person will
be notified via email.Tom, I suggest you start your research with studing the formulae for the process capability factors and understanding what they mean. You won’t learn anything from asking people for their Cp and Cpk values because they will have numerous examples in each company and these figures can range anywhere in the spectrum from positive to negative numbers for CpK’s and from intergers to decimals for Cp’s.
0July 24, 2002 at 10:35 pm #77567Re: CpK values. If you have a process that has a unilateral specification limit (e.g. LSL = 0 USL = 10 – and you want to move the mean as close to zero as possible), is CpK or CpU a good indicator of capability? If not, are there well accepted capability tests to use for unilateral spec limits? All responses are appreciated.
0July 25, 2002 at 1:28 pm #77594
GabrielParticipant@Gabriel Include @Gabriel in your post and this person will
be notified via email.James:
For onesided tolerances Cpk (CpU or CpL deppending on which side) is a good indicator of capability. Just take into account that this kind of characteristics are ussually not normally distributed. So it would be more correct to calculate Cpk based on percentiles or on formulas for the propper distribution that fits the porcess. However, many times the standard formula for Cpk is used anyway. It is still a good indicator of the capability, specially if you use it to compare the process with itself along the time or to asses the efectiveness of an improvement. Only that the PPM based on this Cpk may be unrealistic if the process distribution is far from the normal distribution.
Specifing LSL=0 USL=10 can be confusing. This does not seem, at first sight, a unilateral specification limit. A value below 0 would also be out of tolerance. For example, you can specify that difference in the quantity of CocaCola in the bottle with respect to the labelled quantity shall be betwenn 0 and 10ml (shall not be below the indicated and shall not exceed it by more than 10ml). Of course, if zero is a physical limitation, it will be a one sided tolerance by fact, just because it is impossible to get less than zero. Examples of that are ovality, roughness, delivery time, number of errors. But in those cases you specify 10 Max rather than 0 – 10. One sided tolerances can also be on the other side (like 10 Min), for example the strength of a cable.0July 25, 2002 at 1:44 pm #77598
James AParticipant@JamesA Include @JamesA in your post and this person will
be notified via email.I agree with Gabriel – especially the part about LSL=0 and USL=10 being confusing if you don’t really mean it. If you get run over by a bus or visiting alien spacecraft then your successor needs to be able to see your logic easily in any of your calculations.
If you have unilateral tolerancing, stick with (USLMean)/3sd or (MeanLSL)/3sd depending on your case. Stick with convention – it’s good practice.
Just a suggestion.
James A0July 25, 2002 at 2:30 pm #77606
Mike CarnellParticipant@MikeCarnell Include @MikeCarnell in your post and this person will
be notified via email.Tom,
First I can’t imagine that a study of Cp and Cpk numbers can be involved enough to constitute a study. In the overall sceme of useful tools, it is Industrial Tourism. Unless it is based on LT data it is a snapshot of where you have been – just like the pictures from last summers vacation. In most instances not much help in problem resolution but it is common ground for communication
I agree with Joe’s answer about just asking for numbers. If you don’t understand the amount of data used and the measurement system (MSA) used to generate the data you have a pretty useless number. MSA can drive the number all over the map.
Good luck.0July 25, 2002 at 3:28 pm #77608Hey JoyIts Jilly here. Remember me! I agree with everything you say. Sort out the measurment system first? anyway…….its a load of old cobblers that Cpk stuff. Who uses it anyway.LoveJugsy
0April 30, 2003 at 11:56 pm #85411
Carlos DíazParticipant@CarlosDíaz Include @CarlosDíaz in your post and this person will
be notified via email.How a negative () Cpk can be explained? and how it can be represented in a histogram where Cp=4.4, Cpl= 0.03, and Cpu=8.8?
0May 1, 2003 at 2:45 am #85413
GabrielParticipant@Gabriel Include @Gabriel in your post and this person will
be notified via email.You have a negative Cpk whenever the process average falls outside the specification limits. In your case, the process average is 0.09 sigmas lower than the LSL and 26.4 sigmas lower than the USL. Your variation is very small compared with the tolerance, but it is so ofcenter that more than half of your histogram is below the LSL.
0May 6, 2004 at 12:26 pm #99820
Harvey C. SmithParticipant@HarveyC.Smith Include @HarveyC.Smith in your post and this person will
be notified via email.As Dr. Deming used to say arbitrary numerical goals are worthless.. If you focus on inproving the process and reducing variation the Indices naturally will improve.
0March 10, 2005 at 9:47 am #116185
Venkat Rao KesanaMember@VenkatRaoKesana Include @VenkatRaoKesana in your post and this person will
be notified via email.How do you calculate the CPk of a skewed process, esp when you are measuring unilateral tolerences.
With thanks & best regards,
venkat r kesana0November 2, 2006 at 2:16 am #146365you wrote this: “The value itself can be thought of as the amount the process (car) can widen before hitting the nearest spec limit (garage door edge).Cpk=1/2 means you’ve crunched nearest the door edge (ouch!) Cpk=1 means you’re just touching the nearest edge Cpk=2 means your width can grow 2 times before touching Cpk=3 means your width can grow 3 times before touching” Ken K.
“The general rule of thumb states that if a Cpk value of a process is less than 1.33, then the process is incapable of producing a repeatable part.” Larry Seibel”
My question: what width you mean?!! Is it the car width, or the free space left on each side of the car (when it is parked in the garage) If so, is the two sides free space, or oneside free space. let us say, the Cpk=2 would it mean we have 2 free spaces more on each side more, or 2 free spaces more total, which would mean 1 free space more on each side. it is a good example to remember, thanks..!!0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.