Quality Control

Instruments to do quality control:

Types of Variations
There are three types of variations in production of piece parts:
(i) Within piece variation: This type of variation in quality parameters varies with in
piece, for example, surface roughness of the different faces of the piece.
(ii) Piece-to-piece variation: This type of variation in quality parameters varies from piece
to piece, for example, change in surface finish due to continuous tool wear.
(iii) Time-to-time variation: This type of variation in quality parameters varies due to
different time of the day, for example, product produced in the morning will be better
than the product produced after launch due to laziness of the worker.

Chart Techniques
Following steps are followed to produce a control chart:
(i) Select the quality characteristics: Selection of quality characteristics is very important
in control chart for variables/attributes. Only one-quality characteristics can be used at a
time in control chart for the variable. But, in case of control charts for attributes, more than
one quality characteristics can be considered to define that the product is defective or not.
Similarly, in C-chart a no. of defects can be considered as count of defects in the products.
(ii) Choose the rational subgroup: There are two ways to select the subgroup samples –
product produced at one instant of time or close to that instant as possible and products
produced over a period of time so that they can represent all the products. On the basis
of subgroup size, there are some empirical judgement as discussed below:
• As the subgroup size increases, the control limits become closer to central value,
which makes the control chart more sensitive.
• As the subgroup size increases, the inspection cost per subgroup increases.
• When destructive testing is used and the item is expensive, a small subgroup size of
2 or 3 is preferred.
• On the basis of statistical distribution, subgroup averages become normal when
subgroup size equals or exceeds 4 even the subgroups are taken from non-normal
population.
• When the subgroup size exceeds from 10 or 12, the range (R-chart) should be replaced
by standard deviation (S-chart).
(iii) Collect the data: After deciding the subgroup size, data should be collected to prepare
the control chart.
(iv) Determine the trial central line and control limits: On the basis of the nature of
data collected, use suitable control chart and determine the trial central line and control
limits to check which subgroup is out of control.
(v) Establish the revised central line and control limits: After finding the subgroup out of
control, establish the revised central line and control limits by discarding those subgroups.
(vi) Achieve the purpose: Finally, achieve the purpose of control chart for which it was
prepared.

Control Charts for Variables

Control charts for variables monitor characteristics that can be measured and have a continuous
scale, such as height, weight, volume, or width. When an item is inspected, the variable being
monitored is measured and recorded. For example, if we produce bolts, thread dimension might
be an important variable. We can take samples of bolts and measure their thread major and minor
diameters. Two of the most commonly used control charts for variables monitor both the central
tendency of the data (the mean) and the variability of the data (either the standard deviation or
the range). Note that each chart monitors a different type of information. When observed values
go outside the control limits, the process is assumed not to be in control. Production is stopped,
and employees attempt to identify the cause of the problem and correct it.

Objectives of Variable Control Charts

There are following objectives of variable control charts:
(i) To achieve quality improvement in products.
(ii) To determine the process capability.
(iii) To provide information for decision with regard to product specification.
(iv) To provide information about the current production process.
(v) To provide information about the decisions taken for currently produced items.

23.4.2 X-bar and R-chart

The X-bar (arithmetic mean) and R (range) control chart is used with variables data when
subgroup or sample size is between 2 and 15; however, this is not a fixed rule. The steps for
constructing this type of control chart are:
Step 1: Determine the data to be collected. Decide what questions about the process you plan to
answer. Refer to the data collection module for information on how this is done.
Step 2: Collect and enter the data by subgroup. A subgroup is made up of variable data that
represents a characteristic of a product produced by a process. The sample size relates
to how large the subgroups are. Enter the individual subgroup measurements in time
sequence in the portion of the data collection section of the control chart labelled
measurements (X1, X2, X3, … Xn).
Step 3: Calculate and enter the average for each subgroup. Use the formula below to calculate
the average (mean) for each subgroup and enter it on the line labelled Average in the
data collection section.

Step 4: Calculate and enter the range for each subgroup. Use the following formula to calculate
the range (R) for each subgroup. Enter the range for each subgroup on the line labelled
Range in the data collection section.
Range = Largest value in each subgroup − Smallest value in each subgroup

Step 5: Calculate the grand mean of the subgroup’s average. The grand mean of the subgroup’s
average (X-bar) becomes the centre line for the control chart, i.e., X chart.
Step 6: Calculate the average of the subgroup ranges. The average of all subgroups becomes the
centre line for the R-chart.

Calculate the UCL and LCL for the averages of the subgroups. At this point, the chart
will look like a run chart. Now, the uniqueness of the control chart becomes evident as
the control limits are calculated. Control limits define the parameters for determining
whether a process is in statistical control. To find the X-bar control limits (Figure 23.1),
use the following formula:
Establish the Revised Control Limits
If some data in X-bar and R-charts are out of control, establish the revised control limits by
discarding the subgroups out of control in the corresponding chart. To find the grand average and
control limits following formula can be used.
Example 23.1: A workshop produced the following data (in millimetre) regarding a product
manufactured on 5th August 2014 during two shifts of six hours each as shown in Table 23.1.
Find the control limits for X-bar and R-chart.
Limitations of Chart for Variables
There are following limitations of control charts for variables:
(a) It cannot be used for the attributes type quality characteristics.
(b) If a product has a large number of variables, it will require individual X-bar and R-chart
for each variable which is not possible. In this case, attributes chart is preferred.
Control Charts for Attributes
Attributes are the characteristics of a product or services which can be measured in terms of
binary
value such as pass–fail, accepted–rejected, good–bad, etc., but it cannot be measured in terms of
numerical values such as weight, height, length, width, etc. There are two types of attributes:
(a) Where measurement are not possible, for example, visually inspected items such as
colour, missing parts, scratches and damage.
(b) Where measurements are possible, but not made because of involvement of cost and
time. For example, diameter of a wire can be measured using micrometer, but for
easiness go–no go gauge is used to confirm the accuracy of the diameter.
Control charts for attributes can be used for both defects and defective analysis. Defects
represent
the individual quality characteristics of the product, but defective represents the entire product.
A defective product may have a number of defects.
There are two different groups of control charts for attributes: one is for defective, i.e. p-chart
which is based on binomial distribution and other one is for defects, i.e. c-chart and u-chart based
on the Poisson distribution.
p-Chart
p-Chart is the most versatile control chart which can be used for single quality characteristics
(like X-bar and R-chart) as well as for multiple quality characteristics at a time to control the
entire product. It can be established for controlling the performance of operator, work centre,
department, plant, etc.
p-Chart Construction for Constant Subgroup Size
Trial centre line and control limits
Revised Control Limits and Centre Line
If some of the fraction defectives of subgroups go beyond the control limits, a revised control
chart
is required. To establish the revised control chart, the subgroups having the fraction defectives
out of control should be discarded in the calculation of revised average fraction defective. A
sample
of p-chart is shown in Figure 23.11.

Example 23.4: Inspection results of electric bulb produced by a company in a month of

25 working days are shown in Table 23.2. Daily production rate is 500 bulbs. Find the control
limits using control charts for defectives.
Lot-by-Lot Acceptance Sampling
Introduction
Lot-by-lot acceptance sampling is the basis to check the quality of products produced at different
time intervals, on different machines, in different plants. Different lots of product are prepared
based on the similarity in the production process. All the products in a lot cannot be inspected
due to costly nature of inspection, destructive testing, repetitive nature of job, lack of automation,
etc. Therefore, a sample(s) from each lot is selected for inspection; if number of defective is less
or equal to prescribed minimum number, then the lot is accepted and otherwise lot is rejected.
The acceptance or rejection of the lot is based on different sampling plans which have been
discussed in the next section of this chapter.

Application of Sampling Plan

There are following applications of acceptance sampling plan:
(a) If the testing process used for the product is destructive, sampling plan is used in place
of 100 per cent inspection.
(b) If the testing process is costly, sampling plan is used in place of 100 per cent inspection.
(c) Sampling plan may be used in a repetitive type of inspection to avoid fatigue or boredom.
(d) When control charts are not available, sampling plan may be used.
(e) When inspection process is not automated, sampling plan may be used.

Advantages of Sampling Plan

There are following advantages of sampling plan over control chart or 100 per cent inspection
process.
(a) It is more economical due to fewer inspections.
(b) There are less handling damages during inspection.
(c) It is most suitable for the destructive type of testing.
(d) There is upgradation of the inspection job from monotonous piece-by-piece to lot-bylot
decisions.
(e) There is stronger motivation for quality improvement due to rejection of entire lot rather
than the return of defectives.
(f) Due to fewer inspectors, it is possible to simplify the recruiting, training and supervising
activities.

Disadvantages/Limitations of Sampling Plan

(a) There are risks of accepting defective lots and rejecting good lots.
(b) More time and efforts are devoted in the planning and documentation of the sampling
plan.
(c) Less information is provided about the product.

Sampling Plan
The initial step of the sampling plan is lot formation. Lot formation can influence the
effectiveness
of the sampling plan. The guidelines for lot formation are given below as:
(a) Lot must be homogeneous, i.e. the all the products in the lot is produced by the same
machine, same operator, same input material, and so on.
(b) Lots should be as large as possible. Since the sample sizes do not increase as rapidly as
lot sizes, a lower inspection cost results with larger lot sizes.
The second step of the sampling plan is sample selection. The basic technique of sample
selection
is to assign a series of numbers to each product in the lot and select the products for sampling
randomly. The random number can be generated manually or using software.

Type of Sampling Plans

There are three types of sampling plans used in the lot-by-lot sampling. These plans are as given
below:
(a) Single sampling (SS) plan
(b) Double sampling (DS) plan
(c) Multiple sampling plan
Single Sampling (SS) Plan
In SS plan, the decision about the lot is taken on the basis of quality of only one sample.
The following are the procedures for SS plan:
(a) A lot size (N) of the product is taken to check the quality.
(b) A sample size (n) is selected randomly from the lot.
(c) If the number of defectives in the sample exceeds the acceptance number (c), the entire
lot is rejected.
(d) If the number of defectives in the sample is equal to or less than the acceptance number,
the entire lot is accepted.

Operating Characteristics Curve for SS Plan

There are following steps for plotting the operating characteristics (OCs) curve for single
sampling Plan:
(a) Assume process average, P0 value between 0–8 per cent. Take at least 7–8, P0 values to
plot smooth curve.
(b) Find the value of probability of acceptance, Pa using Poisson distribution table given in
Appendix III of this book (take cumulative value).

The Poisson distribution is used for sampling plans involving the number of defects or
defects per unit rather than the number of defective parts. It is also used to approximate
the binomial probabilities involving the number of defective parts when the sample (n)
is large and p0 is very small. When n is large and p0 is small, the Poisson distribution
formula may be used to approximate the binomial. Using the Poisson distribution
to calculate probabilities associated with various sampling plans is relatively simple
because the Poisson tables can be used.

Plot the curve between 100P0 and Pa. The nature of the curve is shown in Figure 23.18.
Example 23.9: Draw the OC curve for single sampling plan; given N = 3000; n = 100; c = 2.
Solution:
To draw an OC curve, we have to take some arbitrary value of p0. The value of p0 can be taken
from 0.01 to 0.07 at close interval for smooth curve. The corresponding value of probability of
acceptance Pa (cumulative) can be taken from Poisson distribution table given in the ‘Appendix
4’.
The graph is plotted between 100p0 and Pa. The value of Pa = P2 or less corresponding to np0
is
collected in Table 23.6.

The graph shown in Figure 23.19 is slight deviated from the ideal graph due to lack of
exactness in values of Pa corresponding to different np0.
The decision about the sample is based on one or two samples depend on the acceptance and
rejection number of first and second samples. The following are procedures for DS plan:
(a) A lot size (N) of the product is taken to check the quality.
(b) Two sample sizes (n1 and n2); two acceptance numbers (c1 and c2); and two rejection
numbers (r1, r2) are specified.
(c) Take the first sample of size n1.
(d) If the number of defectives in the first sample does not exceed c1, the lot is accepted and
the second sample is not required.
(e) If the number of defectives (d1) in the first sample exceeds r1, the lot is rejected and the
second sample is not required.
(f) If the number defectives in the first sample are more than c1 but less than r1, a second
sample n2 is selected and inspected.
(g) If the number of defectives in combined first and second sample does not exceed c2, the
lot is accepted.
(h) If the number of defectives in combined first and second sample exceeds c2, the lot is
rejected.
OC Curve for Double Sampling Plan
The following are the steps for plotting the OC curve for double sampling plan:
(a) Assume process average, p0 value between 0 and 8 per cent. Take at least 7–8 p0 values
to plot smooth curve.
(b) Find the value of probability of acceptance, Pa using Poisson distribution table given in
‘Appendix 4’ of this book (take cumulative value).

(c) Plot the curve between 100p0 and Pa. The nature of the curve is shown in Figure 23.21.

Treatment of Reject Lots

There are three important treatments of rejected lots discussed below as:
(a) The rejected lots can be sent to production facilities for rectification by production
personnel. This treatment is not satisfactory, since it defeats the purpose of sampling
inspection and slows production. However, if the products are badly needed, there may
not be another choice.
(b) The rejected lots can be rectified at the consumer’s site by the personnel of either
producer side or consumer side. But in this case, there is a psychological disadvantage
since all the consumer’s personnel are aware that the producer ‘A’ had product rejected.
(c) The rejected lots can be returned to the producer for rectification. This is the only
appropriate course of action, since it results in strong motivation to producer for quality
improvement. Since shipping costs are paid in both directions by the producer for
rejected lots.

Consumer–Producer Relationship

The producer always wants ‘all good lots should be accepted’ and the consumer always wants
‘all defective lots should be rejected’. Only one ideal situation for both is 100 per cent inspection
and the sampling plan for 100 per cent inspection is vertical as shown in Figure 23.28. Thus, a
sampling plan always carries some risks for both producer as well as consumer. The producer’s
risk is represented by a symbol α as the probability of rejection of good lot. The values of α
range from 0.01 to 0.10 but generally, its value is taken as 0.05. α is the probability of rejection
and thus its value cannot be shown on OC curve. The probability of acceptance, i.e. Pa = 1 − α
corresponding to AQL is shown on OC curve, i.e. 95 per cent (Figure 23.28). AQL is acceptable
quality level corresponding to producer’s risk. The consumer’s risk is the probability of
acceptance of defective lots. It is represented by the symbol β. The value of β is used as 0.10.
The probability of acceptance of defective lot corresponding to limiting quality level (LQL) is
shown on an OC curve as Pa = 0.10, i.e. 10 per cent as shown in Figure 23.29.