# Calculating Cpk and Ppk

Video Activity

Join over 3 million cybersecurity professionals advancing their career

Sign up with

Required fields are marked with an *

or

Already have an account? Sign In »

Video Transcription

00:00

Hi, guys. Welcome back. I'm Katherine MacGyver, and this is how to calculate CPK and PPK.

00:07

So where we left off in our last module is that CPK and PPK are what we consider to be the voice of the process or that last element in deciding what it is that you can do is an organization to improve your value for your customer.

00:22

So of the three voice of the voice of the customer, which

00:27

is arguably the most important, the voice of the business and the voice of the process, the voice of the process is the only one that could be formulated Klay derived or quantified at rather than qualitative information or subjective information. So you're experiencing broke.

00:43

So in this module, we're going to go through how you calculate CPK and PPK.

00:49

So when you're doing cpk, so remember, this is what your process is capable of. And this is are you capable next to your customers expectations. So your upper and lower specifications limits. You're going to need to have your sigma or your standard deviation,

01:07

and you're going to need your process mean

01:08

So what you want to look for is, if you look at the graph Boots on the very bottom. You have a very tight distribution curve with lots of space on both your left and your right side to indicate that there is quite a bit of space from your upper and lower specifications limit. Which means

01:26

if you have any variation,

01:29

you're still gonna be meeting your customer's needs. So ultimately CPK really tells us, Can we meet our customers needs

01:38

so to do, to get cpk. Remember, you're gonna use sigma where your standard deviation C P. So if you remember back to the previous module, CP and CPK are the same with the exception that CPK is centralized

01:56

or at your target. So you're centralizing factor.

01:59

So to get CPU, take your upper specifications limit minus your lower specifications limit and you divide it by six. Standard deviations on that gives you that idea of where do you perform? Remember when we're talking about a normal distribution, we're talking about our standard deviation.

02:17

Three plus or minus is going to give us 99% and change.

02:22

We want 99% and change of our process outputs to be within our customer expectations. So cp Upper Upper specifications minus lower specification divided by

02:37

six standard deviations. This will tell us whether or not we can in fact

02:42

meet the customer's requirements. CPK it were. So we're talking about normal depth, normal distribution data that has a target generally the average or the halfway point between your upper and your lower specifications limit.

02:59

So for CPK you due to calculations, you do your process mean, which is your average minus your lower um, Laura Specifications. Limit divided by three times your standard deviation.

03:14

Then you will do your CPU, which is your upper specifications limits. So upper specifications limit minus process mean divided by three times your standard deviation. From that you're going to choose the smallest answer,

03:30

and the reason why is because you are looking to be conservative.

03:35

So the smallest answer is going to be your worst case scenario compared to your best case scenario. So for both CPK and PPK or C, P and PP, for all of our process capability measurements, larger is a better score. So we want to go with smallest

03:53

because that gives us a very realistic and then we can delight everybody

03:58

when we, in fact are larger than we perform.

04:00

PPK and PEOPIE are very similar to the same formulas. Um,

04:09

with the exception that instead of talking about our standard deviation for a sample group or for our customer requirements were talking about population data now And the reason why is because PPK is real time. This is what your your process is actually performing at.

04:29

So we're gonna look at all of the data.

04:30

The formulas air the same pp upper minus lower, divided by six standard deviations.

04:36

Um, you are also going to pick the smallest of the answers where you take, um average minus lower, divided by three standard deviations for your Populli shim data.

04:50

So this is going to be instead of a sample where we have a chunk of it, we're going to be looking at all data that we've collected.

04:59

Same with your upper. You're going to do upper minus process mean divided by three standard deviations. And for both of these, both CPI CPI KPP PPK. You're going to be wanting to select the smallest answer because it is a more conservative answer.

05:15

Larger is better scoring for

05:18

both process capability. So if you have a

05:23

a PPK of three What you are telling us is that you can have three versions of your process operating within your upper and lower control limits, which gives you a lot of room to move around or have variation or have a bad day as compared to, say, a cpk of

05:41

0.5, which means that only half of the time

05:44

you are meeting your customer requirements. So this is an indication of the number of times or the number the amount of variation that is allowed within your process, bike upper and lower control limits. And that gives you kind of your safety buffer for your variation.

06:00

So really,

06:02

in addition to knowing what you can and can't do, um CPK and PPK Air really, really helpful if you're trying to determine if your process is in control.

06:12

So if your CPK equals your PPK, you are more than likely in control. And when we talk about in control, will go through a module later on on process stability. But basically what we are looking for is,

06:25

can we do our process repeatedly? We're looking at PPK. We want to look at population repeatedly and still meet our customer's expectations. So what is the likelihood that we're going to be able to go through this process

06:42

with some every day. Variation. Because, remember, process entitlement

06:46

is your best case on DSO still gives something that doesn't need to be reworked at the end of the day or rejected at the end of the day. Um,

06:56

larger is better scoring on, and what that means is larger values equate to better performance. So if we have 0.5 only half the time we're meeting our customer requirements, if we have three, it means that we can, in fact have three whole versions of that normal distribution curve

07:14

within your upper and lower spec limits.

07:16

What that means for you is there's a lot of opportunity to have variation without it becoming a reject or rework or any of the other negative things that come from not meeting your customers expectations.

07:31

So with that today we went over how to calculate she became PPK, or CPI, on MPP. So the important thing to remember is when you're for you as a greenbelt, you're going to be handling normal distribution data the majority of the time, you're going to want to calculate both the upper and the lower

07:51

and take the smaller size.

07:54

Because that is the most conservative way of viewing this.

07:59

In our next video, we're gonna go over operationally defining value. Add So I look forward to seeing you guys there.

Up Next

Similar Content