Precontrol process control techniques

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Precontrol process control techniques
Message posted by Lance on September 19, 2000 at 12:00 AM (ET)

What is needed in a process to be able to use precontrol? How were the 25% 75% precontrol lines determined?

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Precontrol process control techniques
Message posted by JG on September 19, 2000 at 12:00 AM (ET)

You are talking about control charts.
One generally assumes normal errors or a bell
shaped curve - invoking the central limit theorem.
You assume that you have a normal sampling distribution - assuming that an arithmatic mean
is being monitored. You generally assume that the sample mean is centered on the true mean and has a standard deviation s/sq. rt. of n, etc. and then use the Z distribution to
get the 25% and 75% points. These are z= +/- .675, etc. For further details see the references and links at the following URL (129.2.115.68) .

Re: Precontrol process control techniques
Message posted by DL on September 19, 2000 at 12:00 AM (ET)

check this place

http://www.stats.uwaterloo.ca/~shsteine/articles.html

or in dwizard boletinboard
He said to me:

Re: PreControl charts in continuos process (batch)
--------------------------------------------------------------------------------
Juran𠏋 Quality Handbook, 5th Edition gives these steps as far as pre-control goes:

1. Divide the specification band with PC lines located one-forth of the way from each specification limit. If desired, color
the zones appropriately, red, yellow and green.
2. Start the process.
3. If the first piece is in the red (nonconforming) zone, adjust the process.
4. If the first piece is in the yellow (caution) zone, check the next piece.
5. If the second piece is in the same yellow zone, adjust the process towards center.
6. If the second piece is in the green (good) zone, continue the process and adjust the process only when two pieces in a row
are in the same yellow zone.
7. If two successive pieces are in opposite yellow zones, stop the process and take action to reduce the variability.
8. When five successive pieces fall in the green zone, frequency gaging may start and continue as long as the average number
of checks to an adjustment is 25. While waiting for five pieces in the green zone, if a piece falls in the yellow zone,
restart the count.
9. During frequency gaging, make no process adjustments until a piece exceeds the PC line (yellow zone). If this occurs, check
the next piece and continue as in step six.
10.When the process is adjusted, five successive pieces in the green zone must again be made before returning to frequency
gaging.
11.If the operator checks more than 25 pieces without having to adjust the process, the frequency of checking may be reduced
so that more pieces are produced between checks. If, on the other hand, adjustment is needed before 25 pieces are checked,
the frequency of checking should be increased. An average of 25 checks between adjustment is an indication that the
frequency is correct.

Re: Precontrol process control techniques
Message posted by Phil on September 20, 2000 at 12:00 AM (ET)

Precontrol and process control charts are not the same thing. The Juran excerpt describes the classical precontrol approach by dividing up the tolerance zone with lines at 25% and 75% of the width.

Pre-control charts were originally devised to help machine operators set-up a process for production quickly. It assumes that the process is capable and just needs to be centered. Precontrol charts do not help determine if the process "in-control" or not. Therefore, it is not able to help determine if assignable or common causes are present.

A small number of quality practitioners (definitely a minority) teach using precontrol charts in place of the traditional Shewhart type control charts. I personally substituting precontrol charts is very misguided advice and I am in good company.

Montgomery's book, INTRODUCTION TO STATISTICAL QUALITY CONTROL, also discusses precontrol charts.

Re: Precontrol process control techniques
Message posted by Robert Swick (via 207.3.86.92) on July 25, 2001 at 9:28 AM (ET)

Hello Mr. Niles,
I have been trying to incorporate precontrol within our facility covering our production processes, but I am having trouble doing so. Most of the data that we can collect is attribute data. We are an electronics firm so most of the data we collect is at a go-no go rate. Is there any way that you could provide some information to me on gathering precontrol data on electronics or atleast provide some sort of benchmark to look into?

Thanks,
Bob

Re: Precontrol process control techniques
Message posted by Phil (via 165.247.228.126) on July 26, 2001 at 2:54 AM (ET)

I have never heard of "pre-control" for attribute data. However, maybe "acceptance sampling" will suit your needs. Mil-std 105 or equivalent.

Whose idea was it to try and use pre-control on attribute data....just curious because it is not something generally taught.

Re: Precontrol process control techniques
Message posted by Shan (via 129.37.152.51) on July 26, 2001 at 4:54 PM (ET)

I think the only way you could do any kind of pre-control with attribute data is to actually use ordinal data (assign a score based on quality eg. 1-acceptable 10-severe defect)Based on that, you could do some level up pre-control depending on the quality of the product you're getting.

Re: Precontrol process control techniques
Message posted by Darius (via 200.23.217.10) on July 27, 2001 at 12:06 PM (ET)

PreControl implies try to control a process variable, the attribute by it self does not say what to do.

The basic of the tecnik uses sampling of one unit and if it is in yellow take a second unit to colabore (if its in yellow the second ones or in red zone stop the process and adjust).

In attributes case, I don愒 see any yellow zone, If you are just trying to determine the acceptance of the lot, use Mil Std sampling methods.

Re: Precontrol process control techniques
Message posted by Keith M. Bower (via 12.4.141.8) on October 1, 2001 at 4:38 PM (ET)

Be extremely cautious with PreControl - I've seen many instances in where it has been applied in situations and led to a serious breakdown in quality. There's a paper by Johannes Ledolter in JQT a few years ago - unless the process is HIGHLY capable, the false alarm rate can be very high.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.