hello welcome back my dear friends a very

good morning good evening and good afternoon to all of you ah this is the tqm one ah course

and this is lecture number eleven i am raghunandan sengupta from as you know from ime department

iit kanpur so as we are discussing about kaizen and the implementation how toyota as a company

considered the concept of kaizen and concept of total quality in its overall spheres so

we will continued with the seven tools of quality assurance and how they have changed

accordingly has ah the overall progress of of different technology different services

occurred in different spheres of manufacturing and for different areas of with the economy

a comprehensive statistical process control system uses seven actual concepts are tools

to reduce variability and eliminate waste and if you remember elimination of waste reduction

of of ah variability reduction of total cost negative cost i am i am considering trying

to basically implement the concept of quality are are the different important bullet points

for the three stages of quality if you remember the examples which i just mentioned as a examples

for ah motorola g and and all this things so one is a histogram or the stem and leaf

plot and how they implemented and utilize as statistical process control tools i will

come to that any of the checklist as a check sheet any of the concept of pereto charts

and on the concepts of pereto analysis how they can be utilized any of the cost effect

diagram the trying to analyze the actual ah result based on what are the affects which

are happened earlier you also consider the defect concentration diagrams and how they

can be reduce you consider the scatter diagram control charts this will be discussed separately

in greater details later on so i will come to the control charts of x

bar charts r part charts p charts and and all those things so stem and leaf diagrams

are basically graphical representation of the data in number on in graphs so let there

be a set of of a nine numbers so the nine numbers are basically ah starting from hundred

and one two three four five six seven and so on forth so trance construct a basic stem

and leaf which basically if you see the stem and leaf for of a plant or a tree the main

branch has different type of of stems and leafs coming out so basically the overall

emphasis if you see the numbers one zero are common for all the nine numbers

so we will considered them them as the main stem and the leaves would be basically the

at joint numbers after ten which bacs basically makes the sequence of the series are numbers

so it will be to construct a stem and leaf the numbers are divided into two parts the

stem and the leaf and hence the stem would be ten and the leaf will be one two three

four till nine in case say for example it was ah eleven twelve thirteen fourteen till

nineteen so obviously we will have the stem and leaf accordingly it could have been say

for example numbers are one zero zero one one zero zero two till one zero zero nine

so in that case the stem would be hundred and the leaf would be again as mentioned here

one two three four till nine so generally in the number of stem should be less and obviously

the leaf would be the corresponding ask or numbers based on which you are trying to work

so this ah nag nag table which is three point one and all these things have been taken if

you remember i mentioned time again when we were discussing are taken from the montgomerys

book of of statistical process control and the concepts there in so here in table three

point one we consider the cycle timing on in days to pay employee health insurance claims

so the claims are if they are one the days claims days claims are given accordingly

so claims one till till say for example forty are given and all that dates are given or

days are given based on which the claims can be ah settle so if you do the stem and leaf

display so the days are the stem and the leaf are number of days are there in the overall

things so you have a leaf unit of one so if it is say for example leaf of three unit is

one and based on that you have the the numbers which are given in in figure three point one

so that will give you how the idea of the stem and leaf is done above example is ordered

which makes easy to find out the percentile so here if you remember the claim are given

from one to forty so if they are jumbled obviously have to make sequence of how the the claims

are are being done base with on the number of days so the hundred hundred kth percentile

is a value such that at least hundred or k percent of the data values are at or below

the value and at least hundred one minus k percentage of those values would be basically

be ah at or above the value so basically if you are considering fifty percent fifty percent

will basically have half of the of the data half of the probability if you remember the

probability part i mention quietly in detail when we are considering the cdf and the pdf

for the normal distribution so there fifty percent which is the mean or

the median or the mode for normal description it will imply that fifty percent of data is

on to the left on the mean on the mean on the mode and the rest fifty would be on right

hand side so if you looking from your side so obviously different distributions are as

they are they are use would have different medians and different mean values also but

the uniqueness or the on the specialty of normal distribution is because the mean median

mode are the same values so in in histogram the concept is they are continuous data is

divided into intervals which are called the class interval cells are bins and obviously

the class class intervals are done in such a way they are of equal ah hum ah diss no

i wont use the word dispersion they are of equal breath

so if you are trying basically find out the number of students who have see for example

number of have say family members between see for example two two four then another

would be for example from five to seven so obviously in the bins or the intervals are

two in number so you will basically find out number of people who are there with that family

number and note down and add up the whole value to find out the total number of the

frequency the number of bins does not have any any agreed

rules but any any check taken be one of the heuristic is basically as given here where

the number of bins is h is equal to one plus log to base nepeden ah to the binary base

two n with n is the sample size and h is the number of bins which you want to basically

have so obviously they they without going to the

rules which would basically depend on what is your experience and based on which you

are trying to basically divide the the data into bins and try to understand what type

of different type of distributions would be use and what analysis we can do from the data

such that it will give you some idea about the from the quality perspective if you remember

this is not to do with statistics but more to do with the concept of quality and how

you can implement the different type of statistical quality tools in order to improve quality

given a visual impression of the shape the distribution on the of the measures so you

gives you quite a lot of informations and some formation number the inherent variability

in the data can also be found out for the histogram

so the histogram if you look at the bit value and they are equally dispersed on right or

the left obviously you will you will be able to say that the total amount of dispersion

is equal balanced both on to the left or the right or a mean value but now when i am talking

about variability remember that variability or the word which we try to utilize in statistical

terms is his ah variance or standard deviation they are based on the concept what what is

the dispersion based based on fact that how big or how small they are from the mean value

so variability can be considered at different points with respect to median with respect

to mode and all those all those concepts can be utilized but they will you will stick to

the concept of variability based on the mean concept only

histogram is best suited for large data set considering say for example seventy five hundred

and hundred five hundred and ten data points so choy choice of the number of bins becomes

less influential in determining the shape of the distribution so larger data sizes you

will be able to appreciate that they will slowly turn out we are normal distribution

so this is the concept of so called central limit theorem which i am just mentioning it

will be taken up later on as required so again the histograms an example which is

taken from montgomery so they are layer thicknesses and angstrom angdtrom is is is a unit of measurement

on semi conductor way first so the the thicknesses are are given here starting from four thirty

eight forth to four thirty nine and if you see that you have basically three six nine

and ten columns each of them have have the amount of of reading which is noted here so

if you basically with divided into bins and drawn that histogram the histogram is given

as shown here so along the y axis you have basically the frequency and an along the x

axis of the bins of the intervals which are basically to do with the layer thickness in

arms angstrom now one should remember the the the the ah

bins or the interval we have made can be broader or smaller depending on what in the amount

of information which you want so finally you make more better the graph would be and it

will be able to give you a lot of information what is the overall distribution of so called

thickness depending on the number of reading which you have so it will depend more on on

accurate working but it does give you a lot informations from the histogram and whatever

level of accuracy you want that can be increase or decrease depending on the interval length

so i am just considering the interval length efficiency on information can be increased

for statistic point of view using different tools but let us only concentrate on the interval

aspect only so the check sheet would been my collection on the datas measures the overall

idea which are trying to get and the information which you want get check sheet is used to

collect the data in an organized manners specify the type of data to be collected the part

or operation number on the date who are the analyst who collected the data how many different

points were collected at what point of time they were collected and all the information

is important a trail run to validate the check sheet layout and the design can also taken

in order to accurately you find out that the um the data which would be utilize is how

how good or how bad it is is its spurious or is it the collection has been data has

been done in the best possible manner those have to be check

so a check sheet is just given here for information so they are technically just i am giving a

very brief background so the the last two columns is basically has a two thousand three

data and the total amount of so called ah defects for the data which are for year two

thousand two and three and the middle column is basically the yearly data pertaining to

two thousand two and the left most column basically has the entities or what are the

different type of defects they can be supplier parts rust are rusted miscelland weld is there

plains out of limits is there voyage costing is so all of different defects are basically

marked on the left most column so the time oriented summary particularly

variable in looking for trends or other meaningful patterns in order to understand how the overall

process is going on and whether there abrasions based on which you can take immediate actions

to reduce the variability so the pareto chart is basically the next concept as we have discuss

when you are going through the bullet point of discussion is basically the pareto chart

is a frequency distribution or histogram or attributes data arranged by category so you

would basically arrange them ah depending on what level and which areas they make they

can be club use both in measure and analysis step

so you have basically all all if you remember on the leftmost column ah you have the all

different type of defects so they have basically being put along the y axis and the number

of defects are basically marked along the x axis if you see the adc failure they are

a six in number if you see say for example machining number they twenty time in number

so you can understand that what is overall frequency of the error which are occurring

for different defects and then you can basically understand the whole process and take corrective

actions as required so the pareto chart are they do they do not

automatically identify the most important defects but only the most frequency which

are there when the list of defects contains a mixture of those that might have extremely

serious consequences and other of most less importance one or two methods can be utilized

in order to find out what is the level of such some importance which are there for the

defects which can actually adversely affect the overall manufacturing process or the overall

service service forces which you are trying to study using a weighting scheme to modify

the frequency counts can be done so the the more the weights are or more the frequencies

are given a higher weight or later the more recent occurring obviously you can give higher

weight but all obviously it will be that there is some rationale based on which you trying

to give weights so for the time being will consider weights

can we done but will ignore that for analysis and later see at how they can be taken up

so pareto chart as you can discuss the so called leaf stem and leaf diagrams and the

over analysis of the data which are there so it basically parto chart accompany the

frequency pareto chart analysis with the cost of of or exposure of the pareto charts and

what are the what are implement imply cations from the cost perspective so the variations

of the pareto charts are given so component numbers are given and another y axis basic

define or the percentage of components in accuracy located then in the other diagrams

again you have the components and and and all the wrong things which been done like

ah the rust part the plain being be not there costing being not done properly welding being

not done properly so all these are technically marked a long the x axis

and the error frequency on the number of occurrences depending and how you want find out numbers

it can be frequency it related frequency percentage of occurrences so and all these things and

marked along the y axis so if you find out so in in the second diagram you have again

the component on the y axis you have the number of defective components in the next two diagrams

which ah which have have graphs where i am pointing my finder finger you have the wrong

so called wrong numbers are there on the x axis and on the y axis you have the error

frequency and so on and so forth for the forth diagram also

so the next idea which is very heavily used in statistical process quarter total quality

management ideas how to implement them is basically cost and effect diagram once a problem

has been identified we have look for possible sources of problem cause and effect diagrams

are used to analyze and improve steps to find a potential causes and how they can be rectified

also known as ishikawa diagram the methods finds its application in both manufacturing

and service sector in equal proportions several alternative classification methods of the

causes and and caused an effect ideas can be implanted such that very able to increase

the concept of equality which have been taking about for last so called ah ten number of

lectures or eleven no lecture whatever we are trying to complete as of now

so cause and effect diagram so the ah so the four five ms are for my fetching things are

basically machine and technology method or process material which includes raw material

evaluation of of what are the inputs manpower both physical mental and how they have are

being implemented utilize implement in sense that how you are able to utilize the services

of those manpower measurement are basically to do with its inspection some people have

included three more ms which are basically mother nature management and maintenance but

basically the main focus is to try to reduce the the level of of variance in in quality

and try to basically it is the cost also at at add the at wont use the word expense but

in such a way that both both machines or methods or or materials or man power basically are

utilizes in the most efficient an optimal manner

so eight ps will marketings are production services price place promotion people and

personal process physical evidence and packing so obviously they would also have a negative

or positive affect how they being utilize for in order to improve quality the five ss

using service sectors are so now we basically initially we discussed over the manufacturing

sector now they are the service sector so the five ss are surround supply system standard

documentation skills and the scope of work constructing a cause and effect diagram so

how do you do it that so the steps are i just go through them simply very slowly so you

define the problem or or effect to be analyze so what do you want study you basically define

problem where the problem is occurring you try to identify and take and make up plan

that what you want to understand or or or going to the details from the team to perform

in the analysis so obviously they would be co team co team

say for example can be a good engineer can be the workers who working on the machine

and all these things would be there often team would will uncover potential causes through

brainstorm sessions draw the effort box and the central line and basically it see that

where the outline or where the where ah the overall output output means the the overall

system is is producing some output can be services it can be products and we basically

try to find out that where do the product stand with respect to the overall efficiency

of production are overall efficacy of of the quality concept specify the major potential

cause categories and join them as boxes connected to the central line

so basically they would be central line and all the effects would be become so if you

remember the fishbone so that would basically have some idea that how the cause effect concept

can be done so the fishbone there is a means main main backbone and all the bone are connected

to that so basically that that why the cause and effect can be analyst accordingly so rank

the orders and the causes to identify those that seem most likely to impact and take corrective

actions accordingly if you see this is the cause and effect diagram of the fish bone

diagram and example is given so the defects on tanks are there on the main

concern which is the main central line and if you consider the the five ms or or the

the ss whatever there they have been marked according should so they would be for this

thing they can be machines the materials methods which are there on the so where you place

machines materials and thats immaterial so it can on the upper half of of the cause effect

diagram it can be from the lower cause and effect diagram that doesnt matter so they

would measure measurements personals to would be coming in the lower half and how they affect

your total and number of defects in that tanks have different n type of implications so say

for example for the machines it can be one tool too much plays there or the coolant is

not being utilize ah adequately or you are using wrong wrong tool or ambient temperature

outside is very high or there does their humidity so all these things are starting in such a

way that you can at least find out what are the f of the variables which affect the main

component based in which you are trying to analyze the effect defects of the time so

once thus thats done you basically analyze each of them individually and try to find

out in a sense what is the level of important based on which you can analyze all the inputs

which basically makes the cause and effect diagram so effective defect concentration

diagram are there are defect concentration diagrams is a picture of the unit showing

all relevant views the various types of defects are drawn on the picture and the diagram is

analyzed to determine whether the location of the defect on the unit conveys any useful

information about the potential causes of failure or causes of not so efficient performance

of the system so as as i keep repeating not working efficiency

does not means the production is low it can basically mean total cost is high it can basically

mean the overall technology which is being used is very old still we are getting very

bad products or it can be that you had you trying to utilize the new technologies still

we have basically getting back products so useful analyze in the step ah are are only

implementable and applicable as and when we try to implement all the sequence of the fishbone

diagram of the cause effect diagram and the which we basically the ishikawa diagram and

the concept of of say for example the pareto diagram or the leaf and stem diagram we are

basically utilize in a collective sense in order to give the best possible picture or

best possible set of information for the overall working process of of the manufacturing unit

unit or the service sector unit so defects are diagrams i am just considering

from diagram point of view you have basically ah in this diagram which you considering the

the tank ah diagram is given from top view the left side view the back the front the

right side and the defects are basically mark and based on that you find out the number

of of the defect codes are done for a a b c d or or say for example one two three four

depending on how you want to implement the overall system of defect and analyze the defect

of the tank now scatter diagrams are useful plots for

identifying a potential relationship between two variables let us say data is collected

in pairs of two variables x and y say for example like i am given a very simple example

it can be humidity and temperature when you are trying to find out what is the tensile

strength of of product which being manufactured or it can say for example speed of the turning

machine along with what is the temperature of the coolant which is being utilized or

say for example in in in a in a in service sector it may that be when the order of for

food has been made by the customer and whens actually the for food is delivered

so obviously they would be some implement implication based on which you can study two

or more variables in order to understand the overall process which you have try to analyze

the shape of the plot of the y against x will give us some indication the relation between

x and y so scatter diagram are very useful in regression models where you want to predict

forecast depend ah of y depending upon different above set of variables of x which you have

it must we kept in mind that scatter diagram give potentially relations and not causality

so basically they give you the relationship but actually they do not give you any actual

cause and effect concept like say for example a plot rainfall with say for example age of

of the of the children in one district of india

so obviously i will get different type of a plot but what is the actual relationship

what is the cause effect that may not be analyze from the scatter diagram which i have show

you should careful that drawing blind is scatter diagram may not give you the best set of information

which you want so so here you will basically utilize the concept of design of experiment

of a such that we are able to find out what are the actual variables based on which you

can understand where the problem is how the quality improvement can be done how the total

cost can be improved in the sense the overall negative cost can be decreased or the variability

can be decreased and an on all this points are important

so this is the spatter scot scatter plot the scatter diagram indicates a strong positive

correlation between metal recovery and flux so the for the diagram which we studied ah

in few slide back you have basically trying to understand that there are some problems

in the in the tank so how they can be analyzed so you have in the y axis you have the metal

recovery percentage wise and you have the reclaim in in flux in bounce in the x axis

so based on that can understand there is some relationship between x and y and application

of statistical process control improve quality and productivity in cooper plating operations

at a printed circuit board fabrication facility can be done so high levels of defects such

as brittle copper and copper voids word that in the long cycle so can be utilize and reduce

in actual run such that you are able to improve the quality

implement teams was formed for for this example for trying to find out the defect in the tank

and trying to improve the overall working of the process so plating was done for the

tank by the tank operator the operator the manufacturing engineer responsible for the

process and a quality engineer basically got together brainstorm find out the cause and

effect effect diagram analyze the problem and they came up with the solution they would

basically define concentrate on reducing the flow time through the process as mixed missed

delivery targets where were considered to be the most serious obstacle in improving

the productivity of the whole working process of the example which have consider

so an application of statistical process control would be here in measures so excessive downtime

in the on controller that regulated the copper consumption in the plating plant tank was

a major factors in the excessive flow time controller downtime translated directly into

lost production so you will basically analyze brainstorm try to find about what are the

cause and effect diagrams and what are the analysis and basically it will take access

according so this is a little bit cluttered but here the cause and effect diagram for

the controller downtime are analyze where on the on the upper part of the diagrams you

have the reagent replacing paratonic pump failure electric failure are all this thing

are and analyze and how and where the problem occurred are are basically gone dealt in depth

so an application of statistical process control would be the data collection for the controller

downtime was necessary and the cause and effect diagram would basically have this so in the

in the in the leftmost column you would have all the concentration variations what are

the failure causes they were reagent replace replenishing problems was their oxygen controller

had a problems so all these things are analyst the operator basically goes into data that

gives a description on the second last column and and basically analysis more on the actions

which are taken to rec rectify that or remedy the whole whole set of actions which are taken

so based on the data the pareto chart can be made so you again you have on along the

x axis the the different errors of the defects which are there and you analyze the the level

of concentration what are the defects and numbers which are patterned on the y axis

they would give you good feedback that how the process is doing and if improvements are

done based on the feedback or control or the engineer how they can implemented for the

whole system so with this i will end this lecture and continue with by twelfth lecture

the next day have a nice day and thank very much

Thank You Sir for the helpful videos, Can I get the reference of the images, the images are very blurr and the explanations are based on the images