About ZeroRejects

Jim Winings, (see personal page on myspace),  is the Principal of Six Sigma SPC and an independent consultant. ZeroRejects, is a Windows six sigma and SPC, (statistical process control), software application and is very similar to the Six Sigma Public Domain software he wrote for Motorola when employed there in 1986, but greatly modified. That Motorola software was called the A17/78 Supplement Program, and was available to their suppliers to aid them in complying with incoming quality inspection requirements when Motorola started six sigma.

Motorola was the first to win the Malcolm Baldrige Quality Award. The software that he wrote for them directly contributed to that achievement. Cpk-Sigma-Calc allows for six sigma measurements using discrete or attribute data and has a dynamic link to Excel. It generates Yield, Sigma, (Standard Deviation), Cp/Cpk, (process capability), and Defects per Million Opportunies. Cpk-Sigma-Calc is similar to our free DPMO/PPM, Cpk, Sigma, Yield Lookup Table Software.

Motorola needed a uniform way for suppliers to submit material to incoming inspection, also known as VQA, (vendor quality assurance), so Motorola told him to write a program. They made it public domain and sent it to all their suppliers. It was made public domain to allow easy transfer and to allow their suppliers to send it to their suppliers, and so on, and so on. This is what Bill Smith and other decided they needed to get their Six Sigma concepts up and running. Everyone needed to be able to use the same tool at a very low price, free. While this software is not Public Domain or FREE, we feel it is cheap enough to still make that concept work.

ZeroRejects is a very easy to use database application that performs Statistical Process Control, (SPC) for a given data set. The program he wrote for Motorola was the very first six sigma software application anywhere in the world. This software calculates Sigma, (Standard Deviation), Range, Mean, Cp and Cpk, (Process Capability), values and produces a Histogram as well as X-Bar and Range or R-Bar control charts, plus line control charts of Cp and Cpk for Six Sigma or any other quality control applications. It allows the user to select a design margin from 3 to 6 sigma, also know as the 'Target Sigma'.

The program will tell you if you are within the selected design margin and if not, how far you need to move the mean and which way to achieve your goals. It is perfect for where inspectors on the floor or in Vendor Quality Assurance areas that need to do SPC, (Statistical Process Control), and there is not a lot of time to train them on using complex SPC software.

Why Use ZeroRejects

ZeroRejects was the first software to show you graphically and distinctly the shift in your process by plotting a normal curve and a curve over your data. This shift is also called the design margin or target sigma. Click here to see an actual distribution on our features page.

This chart has the design margin set at 4.5 sigma. We have a 4.5 design margin as long as the Lower Specification Limit, (LSL), is less than -4.5 sigma and the Upper Specification Limit, (USL), is greater than +4.5 sigma as it is in this example. That is step one. The next step that must be true is that the mean, (plotted in the middle of the chart), or X-bar Actual +/- 3 sigma, which is the width of the bell shape curve, is inside our +/-4.5 sigma target. And as you can see the red bell shaped curve is inside of the target sigma. This shows your actual shift and if it is 1.5 sigma or not.

What you can also tell just with a quick glance is that the difference between the specification nominal, which would be the mean of the specification, and the X Bar actual has about 1.25 sigma shift lower. Now the specification of this hypothetical part is quite wide so it is easy to obtain such a design margin. If your Lower Specification Limit or Upper Specification Limit is inside of the +/-4.5 sigma design margin, you can not obtain it.

Of course a Six Sigma design should use the X Bar and Range control charts in control rules according to WECO, (Western Electric Company), rules.

With ZeroRejects you can change the design margin so that you can see where your parts actually are. You can also increase it as your process improves, so you can set milestones to get to Six Sigma. And it really is the easiest to use Statistical Process Control, (SPC), package so training for quality inspectors is cost effective.

ZeroRejects and Statistical Process Control,(SPC)

The first step for creating control charts for statistical process control is to collect a sample of data from the process to be monitored using a quality inspection process. This is required to establish upper and/or lower control limits for the X BAR and RANGE charts. You can select how many pieces you wish to collect. Quality books consulted recommend a sample size of 125 to establish control limits. Deciding the correct sample size is a science all by itself and too complex for the scope of this document. But it should be related to a acceptable quality level, (AQL), and a rejectable quality level, (RQL), and usually uses a method called systematic sampling.

In 1986 Motorola required 75 samples to establish the control limits. You can use as little as 25. This is the first step for most quality programs using SPC.

Once you have entered the initial pieces from the sample, (above), you can decide how often to collect additional samples. Once an hour, once a week, etc. You measure additional pieces, (x2-x5 also known as observations or subgroups using stratified sampling), to check that the process is still within the control limits established with the initial sample size pieces above. These addition pieces are added to the database to produce the charts.

The distribution chart is set up with 12 units of sigma across the bottom of the chart. The actual mean of the data collected is plotted in the middle of the chart. A perfect solid bell shape curve is plotted around the mean of the actual data, (X BAR ACTUAL). A perfect dotted bell shape curve is plotted around the specifications. This allows you to see two bell shaped curves side by side. One around the actual data, and one around the specifications. Instead of looking at the shape of the bell shape curve to estimate the skewness, you simply look at the peeks of the two bell shape curves. This tells you how far and in which direction your process is/has shifted from the design specifications. I find this much easier and more comprehensive that those curves that are not symmetrical.

If your lower specification limit is equal too or less than the lower target sigma value and your upper specification limit is equal to or greater than upper target sigma value, and you can keep you process in this window, you should have very few defects.

The bottom line here is you want to keep the 'tails' of the bell shaped curves inside both the specification limits and the target sigma, (design margins), and the specification limits outside of the bell shaped curves.

WECO, (Western Electric Company), Rules

WECO, (Western Electric Company) rules are shown below. They require zones to be established. For more details about how to create zones, see X-BAR and Range Control Charts and how the zones are used. These rules are quite simple and are used to establish if a process is in control or not. The picture below is from the AT&T Statistical Quality Control Handbook ©1956 Western Electric Co., Inc. These are the original rules from the original source.