home *** CD-ROM | disk | FTP | other *** search
- @
- run1
- WHAT IS A RUN?
-
- RUNS - This is any sequence of points for which a large
- percentage fall on the same side of the center line. This
- definition is vague since many experts disagree on what
- conditions to use. The following criteria are used in
- SPC EXpert to identify a freak pattern:
-
-
- (a) More than N points in a row on the same side of
- the center line - where N is provided by the user.
- The default value for N is 7.
-
- @
- run2
- WHY IS A RUN UNUSUAL?
-
- 1. A chart which is in control should have points randomly scattered
- on both sides of the center line.
-
- 2. A run indicates that some non-random or unnatural causes of
- variation are affecting the process.
-
- @
- freak_point1
- WHAT IS A FREAK POINT?
-
- A freak point is any point on the control chart which
- lies outside of the Upper or Lower (+- 3 sigma) control
- limits.
-
- @
- freak_point2
- WHY IS A FREAK POINT UNUSUAL?
-
- 1. 99.9% of all points in a chart which is in control
- should fall within the Upper and Lower control limits.
-
- 2. A freak point indicates that some non-random or unnatural
- causes of variation are affecting the process.
- @
- freak_point3
- WHAT DOES A FREAK POINT MEAN?
-
- 1. Freak points mean that something very unusual
- has happened in the process.
-
- 2. A freak is produced by an outside unusual cause.
-
- 3. It isn't present for very long.
-
- 4. Freak points are the easiest chart patterns
- to find the assignable cause for.
-
- @
- freak_pattern1
- WHAT IS A FREAK PATTERN?
-
- FREAK PATTERN - This is any sequence of points for which a
- large percentage fall more than a given amount away from the
- center line. This definition is vague since many experts
- disagree on what conditions to use. The following criteria
- are used in SPC EXpert to identify a freak pattern:
-
-
- (a) 2 out of 3 points in a row outside of the 2 sigma limits.
-
- (b) 4 out of 5 points in a row outside of the 1 sigma limits.
-
- @
- freak_pattern2
- WHY IS A FREAK PATTERN UNUSUAL?
-
-
- 1. Very few points should fall outside the Warning Limits.
-
- 2. If your process is in control, less than 5% of the points
- on your chart should fall outside of the 2 sigma (WARNING)
- limits. And less than 32% should fall outside of the
- 1 sigma limits.
-
- @
- shift_up1
- WHAT IS A SHIFT UP?
-
- 1. Several consequetive points tending to be below the
- average followed by several more consequetive points
- tending to be above the average.
-
- 2. Something has acted on the process and changed
- the distribution.
- @
- shift_up2
- WHY IS A SHIFT UP UNUSUAL?
-
- 1. A chart which is in control should have points randomly scattered
- on both sides of the average.
-
- 2. A shift up indicates that some non-random or unnatural causes of
- variation are affecting the process.
- @
- shift_up3
- WHAT DOES A SHIFT UP MEAN?
-
- 1. Something new has been introduced into the process.
-
- 2. Something different from what you've been doing or using before.
-
- 3. This new element changes the whole process. At least for awhile.
-
- 4. You have shifted the center of the process distribution.
-
- @
- shift_down1
- WHAT IS A SHIFT DOWN?
-
- 1. Several consequetive points tending to be above the
- average followed by several more consequetive points
- tending to be below the average.
-
- 2. Something has acted on the process and changed
- the distribution.
-
- @
- shift_down2
- WHY IS A SHIFT DOWN UNUSUAL?
-
- 1. A chart which is in control should have points randomly scattered
- on both sides of the average.
-
- 2. A shift indicates that some non-random or unnatural causes of
- variation are affecting the process.
-
- @
- shift_down3
- WHAT DOES A SHIFT DOWN MEAN?
-
-
- 1. Something new has been introduced into the process.
-
- 2. Something different from what you've been doing or using before.
-
- 3. This new element changes the whole process. At least for awhile.
-
- 4. You have shifted the center of the process distribution.
-
- @
- cyclic1
- WHAT IS A CYCLE?
-
- 1. A sequence of several consecutive points which tend
- over time to regularly rise and fall.
-
- 2. A cycle may have several deviations from this pattern;
- however, the overall tendency is for a regular rise
- and fall pattern.
-
- @
- cyclic2
- WHY IS A CYCLE UNUSUAL?
-
- 1. A chart which is in control should have points which change
- randomly with no particular tendencies.
-
- 2. A cycle indicates that some non-random or unnatural causes
- of variation are affecting the process.
-
- @
- cyclic4
- CYCLIC PATTERN
-
- A cyclic pattern is caused by the same thing happening
- at the same time over and over again. Common causes
- of cyclic patterns include:
-
-
- 1. Defective equipment
-
- 2. Fluctuating electricity
-
- 3. Weather conditions such as humidity, temperature,
- and pressure
-
- 4. Human Factors
-
- @
- cyclic5
- CYCLIC PATTERNS & THE HUMAN FACTOR
-
- 1. The most common cause of cyclic patterns is the human
- factor.
-
- 2. Do not forget to look at the human factor when looking
- for causes.
-
- 3. Shift changes and new hires show up on the charts.
-
- 4. Being tired or unfamiliar with with job can cause some
- pretty strange chart patterns.
-
- @
- stratification1
- WHAT IS STRATIFICATION?
-
- STRATIFICATION - Sometimes referred to as "hugging the center
- line." This is any sequence of points for which a large percentage
- fall less than a given amount away from the center line. This
- definition is vague since many experts disagree on what
- conditions to use. The following criteria is used in SPC EXpert
- to identify a stratification pattern:
-
- (a) N or more points in a row which are within the 1 sigma
- limits - where N is provided by the user. The default
- value for N is 10.
-
- @
- stratification2
- WHY IS STRATIFICATION UNUSUAL?
-
- 1. The points are unnaturally quiet compared to the process
- variability found when calculating the control limits.
-
- 2. If the process is in control the points should be more spread
- out. In fact, you should expect over 31% of your points to
- randomly fall above or below the 1 sigma levels.
-
- @
- stratification3
- GENERAL CAUSES OF STRATIFICATION CAN BE GOOD OR BAD
-
- 1. A common BAD cause of stratification patterns is errors in
- measurement - either
-
- (a) Inadvertant errors in measurement such as switching
- to a less precise measuring device, incorrectly
- rounding off results, or a change in the definition
- of what might constitute a defect/defective.
-
- (b) Intentional errors such as "adjusting" measurements
- to be closer to the target (nominal) value or ignoring
- a certain class of defects/defectives during inspection.
-
- 2. A GOOD cause of a stratification pattern could be that the
- natural variability in the process has been suddenly reduced
- by some outside assignable factors. If this is the case, you
- need to discover what happened to reduce the variability in your
- process and continue doing that. Look for a change that
- occured in the process somewher during or before the points on
- the stratification pattern.
-
- @
- mixture1
- WHAT IS A MIXTURE?
-
- MIXTURE - This is any sequence of points for which a large percentage
- fall more than a given amount away from the center line. This
- definition is vague since many experts disagree on what
- conditions to use. The following criteria are used in SPC EXpert
- to identify a mixture pattern:
-
- (a) N or more points in a row which are beyond the 1 sigma
- limits around the center line - where N can be set by the user.
- The default value for N is 10.
-
- @
- mixture2
- WHY IS A MIXTURE UNUSUAL?
-
- 1. The points are unnaturally distant from the center line.
-
- 2. There is too much variation about the center line.
-
- 3. If the process is in control, more points (about 68%) should
- be within 1 standard deviation of the center line.
- @
- mixture3
- CAUSES OF MIXTURE PATTERNS
-
- 1. A mixture pattern is generally occurs when you sample from
- two different distributions.
-
- 2. For example, a MIXTURE pattern can occur when you sample from
- two different machines or processes and "mix" the measurements
- into one control chart.
-
- @
- inc_trend1
- WHAT IS AN INCREASING TREND?
-
- 1. A sequence of N or more consecutive points which tend
- over time to rise - where N can be defined by the user.
- The default value for N is 7.
-
- 2. An increasing trend may have several downward steps;
- however, the overall tendency is for upward steps.
-
- @
- inc_trend2
- WHY IS AN INCREASING TREND UNUSUAL?
-
- 1. A chart which is in control should have points which change
- randomly - both upwards and downwards - with no particular
- tendencies.
-
- 2. An increasing trend indicates that some non-random or unnatural
- causes of variation are affecting the process.
-
- @
- dec_trend1
- WHAT IS A DECREASING TREND?
-
- 1. A sequence of N or more consecutive points which tend
- over time to fall - where N can be defined by the user.
- The default value for N is 7.
-
- 2. An decreasing trend may have several upward steps;
- however, the overall tendency is for downward steps.
-
- @
- dec_trend2
- WHY IS A DECREASING TREND UNUSUAL?
-
- 1. A chart which is in control should have points which change
- randomly on both upwards and downwards with no particular
- tendencies.
-
- 2. A decreasing trend indicates that some non-random or unnatural
- causes of variation are affecting the process.
-
- @