' Prof. K. S. Jaiswal Head Department of Commerce Former Dean & Director Faculty of Commerce & Management Studies | Mahatma Gandhi Kashi Vidyapith, Varanasi- 02 = [email protected]learning. Any other use for economic / cor ed. The users of the content shall not distribute,“When the quality characteristics of the product are ee in such a manner that they are directly measurable and can be expressed in quantitative units of measurement, then control charts for variables are prepared. For example, length of screw in cm., quality of paper in kgs., age of motor tyre in km., purity of gold in carat, etc. Control charts for variables may be of following three types—(I) Mean Chart (II) Range Chart, and (IID) Standard Deviation Chart. A mean chart provides an ong ic quality averages of the samples drawn rom the production process and the main objective of this chart is to scrutinise whether the averages of quantitative measurements of different samples are in the state of control ornot. ‘The processof construction and interpretation of this chartis as follows :— irst of all, the units are selected at random basis ae Rey the mean of cach sample is calculated by dividing the sum of quantitative measurement of the units (included in sample) by the number of items in the sample (n or sample size) as shown below: __ °X = Sample mean = x Xy 4 Xp + X35 Xn X= or a |‘STATISTICAL QUALITY CONTROL T 41 role ¢ The number of items in cach s; i Ne is Esa] ese cf'a sicinle Iga Nee should be equal, According to Mr. Shewhart in of universe (\1) is known, the same: wor eof central lin lable, then it is drawn on the basis of grand average (X) or mean of sample means which is calculated as follows :— = yx X + x, Xs OE op 1H k= Number of samples iber of formulae for finding Upper Control Limit (UCLg)=¥ +322. vn ‘9, = 8.D. of Population Lower Control Limi ulation is not known: In such a case, the standard deviation of the population is estimated on the basis of mean of the range (R) or mean of the standard deviations (G) of the samples and quality control factors in the following manner : On the basis of Ro - On the basis of 6 = ars % 5 i ; 2 is a quality control factor, which dz is 2 quality control factor which | o,presses ihe mato of 3 to Op. It is also expresses the ratio of R to d,. Its value | ascertained from the table of control can be ascertained from the table of | factors keeping in view the size of sample. control factors keeping in view the size of sample. (Gi) Control limits based on the mean of ranges—These limits ean also be determined with reference to another quality control factor Az, the value of which is obiained by referring to the table of control factors for a given size of samples : UCL = X + Agk LCL=X-A3R [Aik =30) - Gi) Control limits based om the mean of standard deviations —The limits cam also be determined on the basis of mean of standard deviations (@) with reference to quality control factor Ay —a Central Line or CL = 3-5 om. Uclp= p43 2235 422M 35 +05 = 3-55 em. n v9 9; 3x05 LCL;'= p-3 2 = 35-7 = 35-05 van vo 45 cm. Upper Warning Limit (UWL;) = 4+ 1565 7 1-96 x -05 54 6 v9 033 = 3-533 om. Lower Warning Limit (LWLz) = p— 1-96 # 7 1:96 x -05 = 35- Hen v9 )33= 3-467 om. ucL; = 35+ ae 35 +038 = 3538 em. 3x05 LCL, = 3:5--=*=35 --038 = 3-462 cm, vié UWL; = 35 4296205 _ 3.5 4 975 = 3.525 om. _ 1:96 x05 LWL; = 35 =35--025 =3-475 em. vi643 Stanstica, Quauity CONTROL | Vn Vas 2Sigma (Internal Control) Limits : UCLy = t+ 20; =-52 42x 002 =-524 em. LCL} = Jog = 52-2 x -002 = 560m. 2 Sigma (Outer Control) Limits : UCLy= w+ 305 LCL = pt — Jog = 52-3 x -002 = 514.cm, Based on these calculations be prepared as follows : X-CHART “52 +3 «002 = 526m, Reach of R : : e samples were -0811 pectively. Estimate the sta ation of the universe from which these samples were drawn. Solution : 5 i ‘On the basis of & On the basis of & R =i by ame 0811 _ 0263, = 0284 = oga7 = 029 uct a a standard deviation of -002 cm. les of size 4, Means of 10 such random diameters have a mean of 25 em, and Control chart for the means of the random samp) samy 25076 25001 25022 2.4966 «= 2-5040 ‘What the process ever out of control ? 25014441 Stanstics Solution : 5 th Given : = 2543x001 =2:503 cm. = 25-3% 001 =2497 cm, ‘On the basis of abave informations, the mean chart will be prepared as follows : CHART From the =, = chart it is clear that the averages of sample Nos. 1.2.8 and 9 are beyond the limits of control. Thus, the process was out of control in case of these four samples. ‘The following data are related to the mean of 20 samples. Size of each sample is 25. The population standard deviation is 12..Draw the control chart for the mean 26, 33, 19, 27 1 + w xs » é 0 re 8 wm 26-3 «245 18-8STATISTICAL Quauiry ContRot | 45 On the basis of these calculations, the mean control chart will be constructed as follows : rt X-CMART 38 38 a4 20 2a 28 28 20. ‘You may use the following control chart constant : for n'=5, A Solution : K+ AgR 4424-58 x 58247564 = a LCly = X-A,R = 442-58 x 58=40836 740-84 -— On the basis of these calculations, the mean chart will be drawn as follows : R-CHART46 1 Statisnes Mustration 7. pertaining to electric bulbs shows the average life and the range of 12 samples of 5 bulbs each. The data are given in no. of hours. Sample Size Sample Mean Sample Range 1 1,080 410 2 1,390 670 3 1,460 180 4 1,380 320 5 1230 690 6 1.370 450 7 1310 580 8 1,630 350 9 1,580 270 10 1510 660 ul 1,270 440s 12 1,200 raw the three lines of the mean chart on ontrol limits for X-chart,, graph paper, You are given that Ay = +577, +, (Meerut, B.Com., 1999) > Solution : / I B Sample Mean or 3X = 16410 tot 2 Sample Range or JR = 5,130 > Central Line for X-chart or ¥ = x, 16310. 13675 RB a TR 5130 _ ann, R= No Tats Control Limits ¥ Chart UCL = ¥+A,R LCL = the basis of these calculations, Range chart or R-chart is usd to 1 é ow the variability or dispersion of the quali produced by a given process. In simple’ words this chart explains the poaition etveuiel im the range of different samples in comparison to average range. The range of a sample means the difference be\ween the largest and the smallest measurement in the sample. ‘The process of construction of range chart is as follows : qa) ee et ‘Of all, sample range is calculated on the basis of dil mallest measurement of units included in each sample, ie, R=L-S — central line is determined as follows : ~ Rp +R,StansticaL Quauty Contra. | 47 inder different conditions the control limits (a) On the basis of S. D. of Range— UCL, = R+3oR LCLy = R-3or (0) On the basis of S. D. of Population— (i) On the basis of quality control factors dy and dy— UCLa = R+3dyo, ot Rady & LCL, = R-3dyo, or R-3ay£ b [The above options are given because 0, =] a Gi) On the basis of quality control factors Dy and D,— _ UCL=Dyo, LCL_=D,-0, (©) On the basis of quality control factors D; and Dy— UCLy=DyR LCL, =D3R Note The values of quality control factors can be ascertained from the table of quality control factors. 2. IfLCL in R chart goes in negative, itis assumed as Zero, as, Redes Re Chart~ ange chart is the companion chart to mean chart and both may be used simultaneously for adequate analysis of the production process. However, the choice between the X-chart and the R-churt is a managerial problem. X-chart is prepared to i analyse the quality averages of the samples drawn from a given process, whereas R-chart is used to analyse quality dispersions of the samples. In practice, R-chars are constructed, first, If this chart indicates that the dispersion of the quality by the process is out of control, it is better not to prepare X-chart until the quality dispersion is brought under ’ = 0-0026 and given that Ay = 0-58,D =Oand Dy SBS ccm UCL and LCL for X and R Charts. Solution : ~ Limits for X Chart : UCL; LCL; = Limits for R Chart : UCLy = Dyk =2-115 x 0026 =-0055 LCL, = DR =0x-0026=0 Mlustration 9. ‘Draw an R Chart for a part of which 30 samples of size 7 were drawn, and hese the samples ranges was JR = 3 cm. It is known that for the samples of size 7.4 = and d; = 0-833,48 1 Stanstics Solution : “ Given : SR=3, k= 30, dz = 2704, ds = 0-833 Central Line or R= R3 =10 UCL = RB tds x Rich +10 + (3 x 0-833) x (-10/2-704) +10 + (2-499) x (037) = +10 + 092 =-192 LCL = BR -3-dy x Rid, =-10 —-092 = -008 ‘The chast will be drawn on the basis of these caleulations. Tiustration 10. Tha factory 10 samples of 4 units cach were taken, ‘The range of samples are as follows : 6,5, 7, 5, 6, 4, 6 8, 4, 9 Determine the Central Line and the Control Limits for &-Chart. (Given yalue of Dy for size of sample 4 is 0 and of Dz = 4-918, d = 2-059) Solution : Central Lineor R = 2B- 8-6 UCL = Dy oy or Dy x aA x gg = 1433 a. r RL gia = LOL = DyoporDy x 20x55 =0 ‘Tilustration 11. ‘A machine is set to deliver packets of a given weight, 10 samples of size 5 each were recorded, Below are given relevant data :— Sample No. 1 2 3 4 5 6 7 8 9 10 Mean X 6 WwW 15 18 7 14 13 b 7 16 Range R 7 7 4 9 8 7 1 4 ul 5 Calculate the values for the central line and the control limits for mean chart and the range chart and then comment on the state of control. : (Conversion factors for n = 5, are Aj = 0-58; Dy = 0/D, = 2°15) Solution : Del ETT Mean Chart: o Central Line X UCL LCL 162-058 x 7-4 = 11-908 Range Chart Central Line or R= 74 UCL = D,R =2115 «7-45 15-651 J LCL = DR =0x74=0 *‘StanisticaL Quauity ConTROL | 49 Comment : Since all the points in X-chart as well as in R-chart are falling within the patrol limits of these charts, the manufacturing process is in the state of control. It means shat the fluctuations arising are due to random causes only and no assignable causes is i ansible. i) §-D. Cha standar used 7 fulfil the objective of R-ch: (1) Calculation of S.D. of samples—First af all, standard devi samples are calculated (0), 02, 03 ...... 4) separately, (2) Determii tion of Central Line —The mean of standard deviation of all samples iscalculated and this mean determines the value of central line, (.c., Do _ Oj +024 Oy +... OF I measure of dispersion, 5.D. chart can also be + The process of its construction is as follows : ions for different Central Line = a= ; (3) Control Limi (a) If standard deviation of the population (o,,) is known— =, 9, os et UCL = g+3—2 LCL=o-3 —& Vin van (0) On the basis of control factors— UCL = Bo * LOL=Byo [Note : If $.D. of population is known, control limits may be determined by using control factors also, UCL = Bo LCL = Ba S.D. chart is interpreted exactly on the same lines as R-chart, It is worth mentioning thal though theoretically a-chart is more appropriate for, contulling the variability of the process, the difficulty of calculating standard deviati Testricts its practical use. In practice R-chart is more popular and widely uscd because of easy calculations. In case af small samples (which is generally the case of control charts) the range is sufficiently stable. Moreover, the range and standard deviati ae likely to fluctuate together. Due to these reasons X and R-charts are preferred comparison of o-chart. Mlustration 12. Draw a control chart for S. D. relating ta 10 samples, cach of size 8. If the sum of 5.D. of these samples is 50 and the population standard deviation is 3-5, Solution + Given : n=8,k= 10, 30=50,0,=35 Cenval Line = @=%2- 8 5 UCL = 6432 2543 25 42062527 625 Vin v2x8 LOL = 5-3 C= 5-2.625 22.375 Vin On the basis of these calculations the control chart can be constructed. Dlustration 13, dard Quality control is maintained in a factory with the help of mean and suite deviation char, Ten items are chosen in every sample, Bightcen samples in e's Shoten whose S¥ w 595-8 and So = 8-28, Determine the three sigma limits 2! Fate50 | Stanstics © You may use the following control factors for finding 30 Timits : a Ay Bs Bi 10 1.03 0-28 172 Solution : Given: SX = 5958, Yo= 828, k= 18,n= 10 Central Line ‘ Mean Chart S.D. Chart = _ BX _ 5958 = _ Eo_ 8:28 R= ESSE 3 3 = Wea 8B ons Control Limits ‘uch = X+aa _ UCL = Be = 33+] +103 x46 = 17x46 = 33] +4738 = 33-5738 = 7912 LCL -AG LCL = BG = 33:1 —-4738 = 326262 = 0-28 x -46 = +1288 Illustration 14, Ten random samples cach of 6 items were taken Frorn the output of a procesé and means, ranges and standard deviations of ten sub-groups were obtained as under : Sample x R o Sample x R o 1 262 a 3 26 6 19-3 9 43 2 221 i 3:2 7 208 8 44] 3 23:3 2 19 8 254 6 50 4 245 5 24 9 23-2 7 10 5 220 UW S1 10 212 5 17 (a) Determine the central lines for X, R and a charts, and (b) Compute the 30 co limits for the three charts respectively. You may use the following control factors }— n Ag By By Dy Dy 6 O48 0-03 197 0 2004 ‘Solution : ti Central Line ¥-Chart R-Chart @ Chart ep ae eer ne ae g = 2a. 318 Xoo Ma al28 = BRS 64 = os 416 Control Limits UCL = ¥+A.R UCL = DR UCL = Bo = 228 +48 x 64 = 2004 x64 197 x 316 22 = 12826 6-225 LCL LCL = DR LcL= Bo = 0x64 = 039x316 = 19728 =0 = -095 of control limits, the process is.in the state of control. Comment : On the basis c itis out of state control in sample 1 in X -chart. R-chart and o-chart. However,aE ASSIGNMENTS SHORT- TYPE QUESTIONS 1. What is Control Chart ? 2. What are Control Chart for Variables ? LONG -TYPE QUESTIONS Write a brief note on the method of constructing control charts for mean and R chart. How does inspection by variables differ from inspection by attribute ?References Gupta, K.L., “Statistical Quality Control” Navyug Publication (2015) 12" Edition Pp. 40- 50 = Sharma, J.K., “Business Statistics", Pearson Education ® Gupta, C.B., “An Introduction to Statistical Methods” Vikash Publication ™ Earshot L., “Essential Quantitative Methods for Business Management and Finance", Palgrave Levin Rubin, “Statistics for Management" Pearson, New Delhi. .S., “Advanced Statistics", Vailbhay Laxmi Prakashan, Varanasi. .L., “Business Statistics”, Navyug Publication Gupta, C.B. & Gupta, Vijay. “An Introduction to Statistical Methods", S. Chand” Delhi Gupta, S.C. & Gupta, Indra, “Business Statistics” Himalaya Pub. House Delhi Gupta S.P., Statistical Methods, S. Chand & Sons, Delhi Nagor Kailash Nath, “Sankhiyaki Vishlation", Minakashi Publication Shukla & Sahai, “Business Statistics", Sahitya Bhawan Publication, Agra