Cpk vs Ppk: Who Wins?

[Music] this video takes a look at two common process capability indexes CPK in PPK my suppliers just sent you your process capability studied based on a poverty supplies you the requirement is a CPK value of one point three three he's provided you one point four so you're happy everything looks pretty good on this chart everything's within specifications fairly well normally distributed you note that your process capability based on PP k is only one point O seven but he provided you the CPK in your happy you just missed something very important about your suppliers process do you know what that is let's start with a brief review of the two indexes and how they're calculated CPK is a process capability the minimum based on an upper spec and lower spec with the upper spec it's you have respect - the average overall average divided by three times sigma capability based on the lower spec is the average - the lower spec divided by three times sigma in this case this little red sigma is calculated from the average range from AK arrange control chart that's in statistical control it's given by the average range divided by D - or D - is a control chart constant that depends on subgroup size CPK is sometimes called the short term capability because it's based on the within variation within a subgroup PPK which is the another process capability index as essentially the same equations a CPK the only difference is the Greek letter Sigma is replaced by the letter S okay and s is simply the calculated standard deviation is taking all the data and calculating the standard deviation just like you used in Excel if here's the function stdev d PK is sometimes called the long term capability because it includes all the data in the calculation of s so why do we have differences the reason you have differences is simply statistical control if CPK is approximately equal to p PK then the process is in statistical control and that means that calculated standard deviation s is about the same as the estimated standard deviation from arrange control chart if the CPK is significantly different than BBK in your process is not in control and your calculated standard deviation will be significantly different then your estimated standard deviation from a range chart let's take a look at how this happens we're going to have two processes that we use the same data in that using X bar and our chart to analyze the results in process number one here there's a X bar chart on top and you can see that all our points are within the control limits and we don't have any patterns so this process is in statistical control we can predict what it's going to make into the future the same is true with the subgroup range chart everything is within specifications we don't have any pattern and we know what we're going to make into the future and this is the chart down here that we would use to calculate the estimated standard deviation to use in the CPK calculation so if you look at a process capability chart based on process one you'll see that our CPK value is one point one nine now you take a look at PP k and it's one point one four very similar estimated standard deviation based on that range chart is nine point five estimated excuse me the calculate standard deviation is nine point nine again very close so if your process is in control the results are going to be very similar now we reorganize the data for process two and now let's take a look at the x-bar chart you can see that this process has points beyond the limits both above the upper control limit and below the lower control limit it's not in control you don't know what it's gonna make into the future a range chart is still in control though everything's within the specif within the control limits and there are no patterns so we're in control on the range chart but not on the x-bar chart so now what's the process capability analysis going to look like well you can see it here and what you'll see now is that the PPK value is the same as before one point one four standard deviation is the same nine point nine it's because we use the same data it's just a range different but the average and calculated standard deviation are the same regardless of how the data is arranged what really changes is when you look at CPK 2.07 and our estimated standard deviation is 5.5 significantly different than these two all because your process is out of control so the values of CPU and Cpl are significantly different for process 2 and it all boils down to that issue of statistical control when there are significant differences in the standard deviations or CPK or vs. PPK it's a very strong indication that your process is out of control and this table shows us it for process 1 and 2 but a standard deviation of nine point nine eight when it's in control nine point five four very similar it was out of control five point four six very different so the reality is that CPK is a better estimate of potential of your process it represents the best your process can do in the short term and if your process is in statistical control CPK is essentially the same as PPK so you really don't need pbk in that case but if your process is not in control you've got something to work on and in reality both CPK and PPK are pretty well meaningless particularly PPK but if you arrange charters to control CPK can give you a measure of what your process could do if you brought it into statistical control so that's a look at C PK or p PK and who wins [Music]

Channel: SPC for Excel

Views: 15,229

Keywords: Process Capability, Cpk, Ppk, Cpk vs Ppk, SPC for Excel, Process Capability Analysis

Id: -h755pWspMA

Length: 6min 20sec (380 seconds)

Published: Wed Feb 06 2019