Joe Perito

I/Mr? I take it you are referring to an Individuals/moving range chart? By definition, an “I” chart is for individuals of “measured” data, or “variables” data. If you want to plot individuals for “attribute” data, then you need to use a C chart when the sample sizes are consistent, or, a U chart if the sample size changes.

There are many reasons a distribution may be non-normal. Perhaps one of the best known is when drill bits wear down and the holes being drilled gets progressively smaller. This will make a skewed distribution. A left or right truncated distribution may be from an auto-adjusting machine the is running at its adjustment limit. It can’t go any…[Read more]

Paul, here’s the correct info. I verified the facts having just referenced the 3rd and 4th editions of the AIAG MSA manuals: 5.15 is used because it represents 99% of the area under the normal curve. This is stated in the manual. When you use the calculations sheet in the (current) 4th edition it has a footnote that says see the 3rd edition…[Read more]

Bear with me Suresh: A process capability study determines whether a process can meet the customer’s specifications on an ongoing basis. User’s of this methodology usually want the process to have a CpK of 1.33 or higher to be called (statistically) “CAPABLE”. A process with normal variation will produce a bell shaped curve with the process…[Read more]

Bank of America only “thinks” they have a six sigma program.  This is refleted in their hiring contract they sent me last year.  It stated that the six sigma engineer was going to be responsible for selling projects to managers who may not want the projects and that the six sigma engineer would be responsible for completing projests with a m…[Read more]

I don’t buy the concept that Manufacturing is the only department that makes money for a company….and I don’t buy the concept that QA/QC is non value added activities.
I realize these notions are six sigma based and accounting based concepts and are trying to make people think in terms of how money is generated and how you can track in on a…[Read more]

Andy,  sorry for the late response.  I seldom look at this site.  A GR&R is performed on a gauge or test instrument that has a read out.  These devices are used for control on the process or job.   In the QA field, the customer expects that any measuring device that is  used for controlling quality have a GR&R performed on it.  A torque gun has n…[Read more]

At first thought, I’m not even sure I would attempt a GR&R on a torque gun.  A GR&R is performed on measuring devices, gauges, or test instruments.  A torque gun measures nothing.  I would be more inclined to do a GR&R on a torque wrench that is used to measure the breakaway torque on the fasteners that the gun screwed in.  Even this would be dif…[Read more]

Careful, you can end up chasing rabbits with the notion that a 1.5 sigma shift applies to anything at all. Motorola made an observation some years ago that one of their processes had a process whose mean varied by 1.5 sigma at the time the observation was made. That was data, that was then. This does not mean a new science was found or that a…[Read more]

A simple thought is to use the Shewhart techniques for process control. Another thought, processes are prone to “blips”. You may not want the software automatically shutting down or adjusting the process. Hook it up to an alarm if so. Review what set the alarm off and make your own call. Are you doing 100% testing with the device? Have…[Read more]

There are “MANY” processes that are not normal bell shaped curves. Measuring holes from drill bits that are continally wearing down will produce a non bell shaped cureve. Any number of processes could have truncated distributions (non bell shaped)if the machine has auto-adjust software with spec limits built in. Welding results and the…[Read more]

Personally, I feel the rates being charged are a black mark on the character and integrity of the people and organizations doing so. Obviously, the time, materials, and investment in the presentations do not merit the ROI. The realism is that you could make a lot of money if you dedicate the time and effort to addressing the dollar problems in…[Read more]

Both charts are to be monitored for statistical trends and both are to be reacted to in the same manner.  Three bearings of diameters of 1.000, 1.001 and 0.999 might be good if a tolerance allows +/- 0.001.  Three bearings measuring 0.989, 1.011, and 1.000 would be disasterous with their variation… but they have the same average as the first gro…[Read more]

Paul, I’m curious to see what your spreadsheet on Shainin techniques is.  Would you please send me a copy?  Thank you.
 

“A rose by any other name would smell as sweet”. The formulae for control limits and capability factors may be different but the scheme of running a capability study is no different. The only difficulty you run into is that the error rates may be tiny and the sample sizes may have to be huge to get useful data. Example, we set up SPC on a seam…[Read more]

Thanks for the lead. I’ll check them out. We’re located in North Carolina just outside of Charlotte.

Task forces, Quality Improvement Teams, Quality Circles, TQM, QFD, Steering Committees, six sigma… the best results I ever saw was where a manager was placed in charge of the (Quality Circles) and this was his full time job to implement it, make it work, train the people, record the project costs, improvements, and savings, set goals, direct,…[Read more]

Guy, I for one am impressed with your credentials. My company has pulled me off my job to lead a six sigma team for two months and they want to “rent a black belt” for a couple of months. This is not a pemanent job but I’m wondering if you might be interested? I told my boss that I’m not interested in a four week black belt and he agrees. We…[Read more]

You’re not fooling me or any of the other ISO savy people Gary. Just for you I did note that the standard allows other methods. However, the savy people also knows the standard takes the trouble to define the world “Shall” and “Should”… and the section that “is” included in the standards for Receiving Inspection, starts off “shall”. We lead…[Read more]

All sampling plans have AQLs associated to them. C=0 rejects the lot with the first defect it finds based on an extremely high confidence level whereas other sampling plans allow the knowing acceptance of defects. Perhaps Gary would like to explain how six sigma provides zero defect raw materials to the process on an ongoing basis without a…[Read more]