QC Training

Some Notation
x1 =1, x2=3,x3=5 x1 =1, x2=3,x3=5 y1=2,Y2=4,Y3=6

x = 1+3+5 = 9
xy = 1x2+3x4+5x6 = 44

c nc
i 1 5

c c c c c c 5c
i 1

Some Notation

c nc
i 1 5

c c c c c c 5c
i 1

cXi c Xi
i 1 i 1

Equation
y = f(x)
20

Y=2x + 10
15

Y = 3x + 5

10

10

 !3 = 3 x 2 x 1

DATA Type
Data

Nominal

Ordinal

Interval

Ratio

Levels of DATA measurement

RATIO
Known distance between any two points
Equality and greater than Relationship Known distance between any two points Equality and greater than Relationship Ranking

INTERVAL

NOMINAL

Equality and greater than Relationship

Unique Classification

ORDINAL

Equality Relationship

NOMINAL DATA
Nominal measurement is the weakest measurement technique Here, numbers or other symbols are used to describe an item or characteristics
Payslip by Emp no Student by roll no etc.

Ordinal (Rank) Data type

Corporate hierarchy like GM, AGM, DGM etc. Satisfactory, Good, excellent etc. Any preference whatsoever

Interval
Data with ordinal properties
>, <, = etc

Distance between two data items can be measured

The diff between 780 and 80 0C is 2 deg C etc

Decision makers can precisely determine the difference between two measurements President>Vice president>Manager cannot be an interval unless we could measure the difference the various title using some equal scale.

RATIO
Have all the characteristics of interval data but that also have a unique or true zero point (where zero means none) are called ratio data. It is the highest level of measurement
Weight of a packet etc. this has unique zero point. Distance , time, money etc.

Data Sources
Internal

External

Alternative methods of obtaining data

Observation Personal interviews Telephone interviews Mail questionnaires

STATISTICS
-a brief introductioin

Some terms
Population
Finite infinite

Sample Statistical inferences Concept of error Various distributions

Population and sample

Population
A collection of all the items or observations of interest to the decision makers
Finite: countable nos of items
Account receivable

Infinite: uncountable nos of items

Physical inventory

Sample Subset of population

Population Vs sample

Measures of location
Mean Median Mode
A set of data may have two or more mode

Measures of spread or dispersion

Range: max-min value Variance Standard deviation

t-distribution
If the population from which the small sample is selected is normally distributed then t-distribution used t-distribution is symmetrical, ranging from infinity to + infinity

Normal dist and sigma

1 2

Different Z with same mean

Z vs t-distribution

Variance vs std dev

x =24/6 =4
X (no of car sold)

(x-x)
8-4 =4 1-4 =-3 1-4 =-3 5-4 =1 6-4 =2 3-4 =-1 (x-x) = 0

(x-x)2
16 9 9 1 2 1 = 38

2x = 38/6 = 6.3 car squared x = 2.5 car

We can not visualise what car squared means. An important reason for using the standard deviation as a measure of spread is that it is expressed in the same units as the mean

8 1 1 5 6 3 X =24

Std vs Variance

Standard deviation = SQRT ( Variance)

Continuous or Variable data

Yes

Is each data point a natural subgroup? Or Are data gathered infrequently?

Yes

Are data normal?

No

I- chart MA/MR X-bar and R or x-bar and s

No Yes

The data are counted

Attribute data

Yes

Are defective items counted?

No

Can sample size vary?

Yes

p-chart np-chart u-chart

No Yes

The number of defects are counted?

Can sample size vary?

No

c-chart

Control chart decision tree

Decision for control chart

Number defective'np'chart. Percentage defective 'p' chart. Number of defects 'c' chart. Number of defect per unit 'u' chart.

Control Chart
CL WL 2 1

WL CL

Indication of Assignable cause

A single point outside the control limits Two out of three successive points are on the same side of the centerline and farther than 2sigma from it. Four out of five successive points are on the same side of the centerline and farther than 1sigma from it. A run of eight in a row are on the same side of the centerline

Interpretation of Control Chart

Interpreting control charts correctly is necessary to accurately determine variation in a process. Using the sample data, variation is determined based on the probability of producing acceptable product. Several rules for interpreting control charts are widely known and should be used to determine whether the process is out of control:

The Eight rules

1. 2. 3. 4. Data point is beyond control limits. Nine points in a row on a side of centerline. Six points in a row increasing or decreasing. Fourteen points in a row alternating up and down. 5. Two out of three points in a row beyond two sigma. 6. Four out of five points in a row beyond one sigma. 7. Fifteen points in a row within one sigma. 8. Eight points in a row on both sides of the centerline within two sigma.

First, let us assume that the process conditions are such that using a subgroup size of n = 1 is not a logical requirement. Under these conditions many SPC texts recommend that the data be subgrouped so that the central limit theorem applies to the subgroup means and hence the means will be approximately Normally distributed. These authors argue that the individuals control chart is very sensitive to nonnormality and hence the Type I error may be significantly different than what is expected. They conclude that the individuals control chart should be avoided unless the data is approximately Normal or can be transformed into an approximately Normal distribution [Montgomery, 2001]. On the other hand Dr. Wheeler says that control charts work irrespective of the distribution, so there is no reason not to use them for individuals data. I am going to take the liberty to suggest that when Dr. Wheeler says that they work he means that for all practical purposes differences in the false alarm rate (i.e., Type I error) do not significantly compromise the adequacy of the control procedure [Wheeler, 1995]. The logic supporting this argument goes back to Shewhart who based his conclusions on the Camp-Meidell theorem, which states if the distribution of X is unimodal (i.e., monotonically decreasing on each side of the mode) and the mode equals the mean, then the probability that X will deviate from its mean by more than k times its standard deviation is less than or equal to 1/(2.25k2)

Now here is my perspective of the issue. In contrast to Dr. Montgomery's view I feel that the individuals control chart is to be preferred over the mean chart because the mean chart obscures the vision of the practitioner by preventing them from seeing what the process actually looks like. That is, using descriptive statistics (i.e., the mean) to characterize the behavior of the process rather than the actual data results in a loss of information, and game theory teaches us that the best decisions usually result from having the most information. I believe it is critically important for the analyst to have an accurate picture of the process, and the only way to do this is to look at the actual distribution of the individual observations.

The practitioner should be concerned about the false alarm rate, if the cost of failure is high relative to the cost of inspection or if too many false signals seem to be undermining the confidence of the operators in the program. So, if you are concerned about the false alarm rate, you can always model the observed data distribution and use probability control limits to get more accurate results. Remember that a picture is worth a thousand words, and an accurate picture is worth ten thousand! - Don't throw away information!

Statistical Basis of the Control Chart

H0: = 74 H1: 74

Statistical Basis of the Control Chart

Types the control chart Variables Control Charts These charts are applied to data that follow a continuous distribution (measurement data). X-bar chart, R chart, S chart, Attributes Control Charts These charts are applied to data that follow a discrete distribution. p chart, c chart, u chart,

Choice of Control Limits

General model of a control chart

UCL W L W Center Line W LCL W L W

where L = distance of the control limit from the center line W = mean of the sample statistic, w. W = standard deviation of the statistic, w.

Choice of Control Limits

99.7% of the Data

If approximately 99.7% of the data lies within 3 of the mean (i.e., 99.7% of the data should lie within the control limits), then 1 0.997 = 0.003 or 0.3% of the data can fall outside 3 (or 0.3% of the data lies outside the control limits). (Actually, we should use the more exact value 0.0027) 0.0027 is the probability of a Type I error or a false alarm in this situation.

Choice of Control Limits

Three-Sigma Limits

The use of 3-sigma limits generally gives good results in practice. If the distribution of the quality characteristic is reasonably well approximated by the normal distribution, then the use of 3-sigma limits is applicable. These limits are often referred to as action limits.

Choice of Control Limits

Warning Limits on Control Charts Warning limits (if used) are typically set at 2 standard deviations from the mean. If one or more points fall between the warning limits and the control limits, or close to the warning limits the process may not be operating properly. Good thing: warning limits often increase the sensitivity of the control chart. Bad thing: warning limits could result in an increased risk of false alarms.

Quality control during operation

The Quality Trilogy

Quality Planning
40 Quality control during operation

Sporadic spike
Original zone of Quality Control

Cost of poor quality

20

operation

Begin

Chronic waste

New zone of Quality Control

Quality Improvement
Time

0 Quality improvement

Quality Planning

Determine who the customers are Determine the need of the customers Develop product features responsible for customer needs Develop processes able to product feature Transfer the plan to operating forces

Quality Control

Evaluate actual product performance Compare actual performance to product goals Act on the difference

Quality Improvement (During operation)

Establish infrastructure Identify improvement projects Establish project teams Provide the teams with resources, training and motivation to

Diagnose the causes Stimulate remedies Establish control to hold the gain

Joseph Juran is an internationally acclaimed quality guru, similar to Edwards Deming, strongly influencing Japanese manufacturing practices. Joseph Jurans belief that quality does not happen by accident gave rise to the quality trilogy: Quality planning Quality control Quality improvement

The key steps in implementing company-wide strategic goals are: Identify customers and their needs both internal and external and work to meet those needs Create measures of quality, establish optimal quality goals and organise to meet them. Create processes capable of meeting quality goals in real operating conditions. In the 1980s Joseph Juran recognised that the common approach to total quality management - quality awareness campaigns and slogans - was not effective as they did not have substance, and there is no short cut to quality. He believes quality must start at the top, irritating senior managers who believe training is for junior workers.

Indication of Assignable cause

Fifteen points in a row within 1-sigma of the centerline Six points in a row steadily increasing or decreasing Fourteen points in a row alternating up or decreasing Eight points in a row all farther than 1 sigma from the centerline, on either side A jump of 4-sigma

Something to consider

Control limits are not specification limits. In fact specification limits should never be placed on a control chart. Specifications reflect customer requirements, whereas control limits reflect the historical performance of the process.

Control Chart
UClx X A2 R LClx X A2 R

UClR D4 R UClR D3 R

Calculate the control limit for x-bar

Size of sub group (n)

A2

D3

D4

d2

3/d2

2 3 4 5

1.880 1.023 0.729 0.577

3.267 2.575 2.282 2.115

1.128 1.693 2.059 2.326

2.659 1.772 1.457 1.290

0.483

2.004

2.534

1.184