|Chance Causes||Assignable Causes|
|(i) Consist of many individual causes.||Consists of one or just a few individual causes.|
|(ii) Any one chance causes results in only a minute amount of variation. (However, many of the chance causes act simultaneously so that the total amount of chance variation is substantial).||Any one assignable cause can result in a large amount of variation.|
|(lit) Some typical chance causes of variation are :
(a) Slight variations in raw material (though within
the specifications) (6) Slight vibration of machine (c) Lack of human perfection in reading instruments
and setting controls.
|Some typical assignable causes of variation are : Batch of defective raw material
Faulty set up Untrained operator
|(jy) As a practical matter, chance variation cannot economically be eliminated from a process.||The presence of assignable variation can be detected and action to eliminate the causes is usually justified.|
Approach for Control. If samples from process or incoming lots continue to follow this predicted pattern, the process may be said to be under control, or lot may be presumed to be
of quality conforming to process standard. Alternatively, we may infer that the process has undergone change, or that the lot does not conform to attainable standard of quality. Attempts are made to retain this change if it is desirable (i.e. revise standard also) or to avoid it, if it is undesirable. Of course there is a risk of taking wrong decisions which may be restricted to desired degree.
Results of samples may be graphically presented in the form of a chart with Control limits based on process capability (dealt with in next chapters) to detect any undesirable change in early stages (through any shift in the expected pattern) to enable preventive measures during production. This avoids wastage of resources like material, efforts of operators and technicians, machine, etc. It is important that the order of production is recognised and all allied information about production such as product size, delivery time, operator, machine, etc. are maintained on the graph.
A process under statistical control, need not yield products informing to specifications. An incapable process vis-a-vis its specifications should be viewed on its merit.
The main philosophy of Quality Control is prevention of defects during production. This is best achieved by the operator himself. This aims at control of quality by the operator rather than control of the operator for quality. In this context the concept of Self Quality Control attains top priority. Necessary education and allied adequate services and facilities have to be provided for. These activities will cost something, but the absence of these will ultimately cost the company much more. It is essential to keep records elaborately to enable its classification according to all possible factors that influence the quality.
Approach for Breakthrough. Contrary to control aimed at maintaining the attainable process standards, “Breakthrough” aims at improving existing standards of performance; A reference may be made to Figs. 18.12 and 18.13.
As e. fn’.-ii ..rep, enumeration of all causes or sources having impact on the totality of outcome, namely, cost, production, productivity and quality including reliability it made. This is done through a *brain storming^ session involving all concerned personnel from design, production, inspection, etc. from all levels including operators. This is conveniently presented in the form of’Cause and Effect Diagram’. This is also known as ‘Ishikawa Diagram’ after the name of its introducer or ‘Fish-Bone Diagram’ by virtue of its appearance.
Figs. 18.12 and 18.13 show typical examples of Irshikawa Diagram of Welding output Quality* and ‘Shift’ in castings. From a study of the diagrams it is possible to know the diversity of factors influencing the quality characteristic of interest. Accordingly appropriate planning for further study for a break-through approach can be organised.
A—The difference between the old standard and the new standard is regarded as a chronic ailment which can economically be curved.
Next step involves classifying all the quality faults observed hitherto according to the causes and sources and identifying the vital few (Pareto Principle) among these for incorporating preventive measures that will ensure high ratio of gains accrued to input investment in various forms.
Pareto type curves originally intended to exhibit unequal distribution of wealth can be applied to quality losses in any form. It is experienced that only a few of the many causes or sources account for major portion of the losses in the form of rejections, rework, scrap and even capital
Fig. 18.14. Analogy between distribution of wealth and losses.
locked up in inventory of raw material of semi-finished product or finished products. The curve showing-cumulative per cent of causes or sources or items versus cumulative losses on account of defects or capital blocked is shown in Fig. 18.14.
Consider monthwise data or rejection of stators in winding section. Summary is presented in Table 18.1. It is seen that from among seven different causes of rejections ‘HVF’ alone accounts for about 70% rejections every month. The rejections due to ‘HVF’ and ‘UBC account for more than 80% of the total rejections each month. Thus rejections recur in such a way that a few (a small percentage of) causes account for a large percentage of rejections of stators. -
Table 18.1 Illustrative example of Pareto Analysis of Rejections of stators.
This is a universal phenomenon. In stores, 15 to 20% of the total number of items account for about 85 to 90% of the yearly consumption value and only about 8 to 10 percent of the total number of customers account for about 70 to 80 per cent of yearly turnover. Many such examples can be given. This is a very useful feature for it helps in isolating the VITAL FEW from TRIVIAL MANY which need to be given concentrated attention for ensuring best results.
|Month||Percentage of Rejections due to||% Rejection Total|
Besides specially designed experiments, simple and multiple regression studies and allied techniques including Operations Research may have to be made use of in locating the factors that matter and in estimating the degree by which these need to be altered for effecting desired improvements. It may also be necessary to conduct pilot confirmatory runs before making formal process changes that should be incorporated in the process standard to make it upto date. In many cases it is possible to maintain process details in such a manner as might avoid experimentation. These data collected over a long period relatively form the basis for determining optimum levels for various factors on plant scale itself. For growth and stability of an organisation and in fact for its existence in this, world, it is necessary that it passes through successive cycles of breakthrough and control.