QuikSigma: Process Capability for Six Sigma Projects

In this video, we discuss how to set up and interpret a Process Capability for Six Sigma projects using QuikSigma. Have any questions? Please leave a comment below and we’ll get back to you.




The basic idea a capability study with interval or ratio data is fairly simple and that this, we can have a process behavior chart or the histogram of the data in the process behavior chart and that represents the voice of the process. That’s these blue bars here and the red curve that’s drawn as an approximation to them. We then compare that with the voice of the customer. Well what’s the voice of the customer? Usually, that comes out in spec limits and the basic idea is fairly simple. If we know the spec limits and if we know the mean and standard deviation of our process, then we can tell how much of the process is going to be out here in the tails beyond the spec limit. Now, this particular diagram is setup with the spec limit’s at +2 sigma and -2 sigma. So we know all that we’re going to have two and a half percent of our instances out in this tail, and another two and a half percent out here. So we’re gonna pass about 95 percent of our goods, and fail the other five percent.

So let’s go to the simulator and see what we can find in the way of data. This simulator has an output variable, which we see down here, in this case 363, and we have a set of input variables to the process that we think might be driving it. Well, you can go in and modify this table and set yourself up the designed experiment and determined how much each of these drives the output and that’s a little beyond what we’re going to do here. I’m just going to collect a set of output data and we’re going to analyze that and all I have to do is just to start this and then I’m going to copy this column over into QuikSigma so we can analyze it.

So, I’ve opened capability under baseline and of course I could do the same thing under the tool box if I wanted, it’s available there. I pasted my output variable here. My default is I-MR, which is appropriate. I’ve checked my four Western Electric rules and I probably should have pointed out to you in the simulator that upper spec is 1085 in our lower spec is 565. So now all I need to do is click calculate and this is a very familiar result. We’ve already done behavior chart so here’s my individual and here’s my moving range. Again, I’ve got my histogram here but what’s been added is the specification limits, 565 in 1085. Well, we can just look at that and say wouldn’t it be nice if we could just move that up to center it a little bit and we also can just look at that say if it’s a little too fat to fit nicely between the spec limits, we’ve got a problem. Well, we really don’t have time in this session to go into PPK, CPK, PP, and CP. Those are quality indices that an expert can use to derive information about the strategy for fixing the process. That’s just interpreted for you over here so it’s really easy. It says that we’re making 694 and some odd thousand defective parts-per-million and if I remove the special cause it’s not going to change very much.

As matter fact, the difference between those is just due to statistical error. Okay well, why is that? If we look at this, there is no special cause. This is a stable and predictable process. So removing a special cause isn’t going to do anything for us. Now, what it says is that if we center, we can get down to about that 29,500 and some odd defective parts per million which still isn’t very good and that if we do both theoretically we can get down to this level, which is really just not statistically different from that. So, what we find out is that we need to adjust something to move us to the center, but even that isn’t going to solve the whole problem. We’re probably going to end up redesigning this process to also make it skinnier. So my work is laid out for me.

Return to Blog Posts >>

Leave a Reply

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