Six Sigma: P Charts and Capability

In this lesson, we discuss using p charts and capability to help you with your Six Sigma workflow. You can get your p chart and capability resources here. If you have any questions about how to use p charts, or any other Six Sigma questions, please leave a comment below!

 

 

Transcription

What we’re going to study in this session, is using the p chart and its form of capability to estimate whether your process is stable and predictable, and whether it’s capable of meeting requirements. Now, for this, I’ve hijacked the defects per unit and tortured it and made it do things that it doesn’t naturally do.

What we have here is the number of items that were inspected on a given day. Day one, I inspected a thousand and six items and the number that I rejected, the number that I kicked out, was a hundred and eleven. Those are defectives. This will generate 25 data for me. I’m going to take that data and transfer it over to
QuikSigma and we’ll see how to use a p chart. So here I pasted the 25 data that we got, reversing the order of the columns. This is the number that I looked at each day, and this is the number of defectives that I found.

Now there are four conditions that we have to meet to be able to use a p chart. One is that the numbers in this column have to all the integers. Well, that’s easy, we’ve got that. The other, the second one rather, is that everything in this column has to be able to be classified as having or not having a characteristic. So the characteristic that we’re looking for is that they don’t pass. So the ones that don’t belong to that group do pass. So we know if we’ve got a thousand and six that we’ve looked at, a hundred and eleven were kicked out we know that 895 were good. Now, the third condition is that, while I’m taking any one of these samples, the proportion of defectives, or the proportion with the attribute I’m looking for, has to remain constant. Finally, the fourth condition is that the numbers in this column are independent of each other. In other words, we’re assuming that defectives do not come in clusters. The fact that I got a hundred and seven on this day will not help me figure out that I got 150 on this one and 113 on this.

With all that done, we have to make sure that this box is unchecked. In fact, unless that box is unchecked, we don’t get this subgroups column, and then we’ve got a constant sample size. So I could check that and say, give me a constant sample size of a thousand and I’d be about right. But let’s go back here, since we’ve got that data, and let’s just click calculate. We can see that the process is stable and predictable. That is it staying within the limits? You might notice if you’ve got good eyes that there are little squiggles in the limits. You might wonder, what causes that? Well, let me just go down here and make an exaggerated example. That comes about because your sample size is not constant. Look at that. You see when I reduce that to 1500, or excuse me, to 500, I get a little jiggle in my line. Wow, of course that proportions really bad.

The other thing that I get out of this is that I get an estimate of my capability. What we’re looking at here is the proportion that are defective or, if you want to think about it, percent. If i just look over here, I’ve got about 12.5% that are bad. In one sense, that’s my process capability. I can see that down here, 126,155 average defective parts per million. I can translate that over into a PPK number if I like. PPK of .38. Hey, that’s a great opportunity. Typically, we want our PPKs to be above 1.5 and 2, if we can get it, to get a a Six Sigma process. So lots of opportunity there.

Return to Blog Posts >>

Leave a Reply

Your email address will not be published. Required fields are marked *