ConTRot. CHART Based en the Ueeory of Paobabiltiy and sampling , ShewWhart's cerhal chart provide a powerfil tool of . cliscoveniing amd cesvecting the assignable came of veruation | outside. the *sfable pattern’ ef chamet cause , thus enabling us to stabline and control owr procenses at desived pert formants - tn Industry ene fs faced with two kinds of problem : © Te clack whither He proces fs conforming to standard Bald dousn @ TT improve. the standarel (Lael) and sedeeee veouia bility consistent with cost cemsidnations « Shuohant's conto! chants provicle, an answer to both: Cemhal Chas Ps simple pictorial alytca that detecting unnatuutl pattern of vanfatien fn data sewalting from supiteetise proces’ Suewnartis CONTROL CHART For VARTABLE:- for a chanactenistic which can be measwud quantitedvety Mamy quatity characteristics of a procuct are measwiable and can be expoersed tn specific units of Measwement suck as cUameter of Screw, Orfe of Bulb ele: Such verrable ant conlinusus type and follow Normal probablitty Soro Lihen we deal with measurable quality charasenistics, t necewary to conhol tla central tendemes (average) as well ar clispension (variability ) of the quality characteristic of tls procvd + Than. aw differ control clrantp fos variables for controling mean and vaniabiiity of the process - | @ Chants for K amd R (Range) @charty for Kamd ©(S-D+) \X and R (RANGE) CHARTS :- at UNA RANGE) CHARTS *- No precduch'en process fs penfect emough to produce cll the Hems exactly alfke + Some amount of varvation , in produced itemp, ip tnhenemt in any proctuction schime» “Ms variation ts fotally ef Mumenous cheractentstics of tle procw viz FW material, machine setting and handling ett We gemenally use combo! clot for mean (XK chant ) te Cenkol tle mean quality Level ov procus mean: Expeniemee seveals the Protrs VerUability also affecis procuct quality. wWhin variation in the poocuction proccss Ts high. that ip. the proclueed dems have a wide Senge. of values then tle quality of the procluet ip assumed to be poor: So for improvement fn tle quality of procluct we must seduce tle variability : Steps For X Ano R Cuarrs:. ERS POR A AND R_Cuarts: 5 Measurement + - Any method of measwrememt has ls own veuiab/itty « Evrove im measwement cam efthia enten into data“o use of folty imshumemk, flack of clan—cut clefinatien of quahity clemacterss . method of taking Measuruments ° Sines lla conctusion drawn foom contsol chev fave bvoacll besed on lhe veviability a, Measuncmemnt as well as Ube veruabllity Fa tle quality being meascud. él is impoateat sthat the “mistake fn seacding ‘measu'ument Pishuments oF Crreep Im ducercling Ata shoud be minimised as to drat valid concltuion” acm control chert « 2) Seleetion of Samples ov Sub =Jroups :- Jn onclea fo make the combrel chant amalysi effective. 1! i erential, to pay uc segrrt to the Sealional seleciion of Ms samples o¥ sub gaeupp.3) Calculation of X and R for each Sub-group -- am» be the measwuments on the th sample let xP jena Cie12 ok) es The mean Ky, tle sraonge Rr and SD sj fer the i) sample cw given by I : oe ee oo r . ea z Xp) Ree max RXy - il Xi sP= dE (xj -%) (12,228) Next we find X Rand 5, the avenge. of sample mean, om ple Somes andl sample standand deviations (ims ease onsd Re eRi , te 4) Setting of Conbool Limits +. Jt fs welt known that &Y o fs the process slamdard deviation (otamdand drviation of the universe from which samples au daken), then the standand essen of sample mean is O/ 5 where D fs scumple she , SE (Ke) = lpm (bz... k) coe knoe that Also from 6 the sampling dsbibuion of van E(R)=d,.0 "dy =censlant (depend on . sample sinc (my) hus am estimale of © can be obleiined from R by Relaliem- Redo 2» @.F ; 6) /4, Alao X gives am unbinsed estimede of populalion Mean ¢/ E(X) +4 FPR) = tea nl Anse@ Conbol Limits for X-Chast - easel whem stamdand au known (given Ly om) Tre 3-0 Conbrol Dimfts for X Chant cum given by ~ E(K) + 3-SE(X) ut 3. vn Hot Avo (A= Vm) if aS wove kmown oF specified values of eH! ©, saspecticely than , UCLA) = M+ ho LCLigys U'-Ac" case 2:- — Standands nol given (Ae ane unknown) Ulu using luis estimates K and & Jwpectively , we. ge tie 3-0 cenbool btmits fos the X-chant cw! eee Rechont ts combucted fer cen kelling sion (vera bil fy ) ef (be precluel + ding R-chad — fp similay to x the varicion in the elisper- he precodn of combue chee + 1) Compute (le. Ree nge ae man Xy = mn x ( ts 1 2..m) for each sample» 2) Compute the mean of tke sample Sranges Bet PRP e 2 (BAR FR) Ki kK@ 3) Computation for Contool Limits: - “the 3-0 comhol Matz for R- chant ane E(R) 430% ECR) Ts estimated by R and ox is estimated foom the suclation Og = dye z dd, we constants ES dye R 2 3° n a depending en 7 UCL (py = EIR) 4 30K = RL3B- RB a, = Re 3d qd, = (14 3) ca UClLipy= Da R Similarly, LCLeey = (1 when o- ts kmousn UCL, = FIR) 4 30g = d,o + 3d,o- = ,43d,)o = Do LCL ipy = EIR) -30y = dy. 0°-3d, 0° =(d,-3d,)o> = Dom n eacl. case (0° kmown o” unknown) Clips FIR) R 4). Conbuttion of Combol Charts for Xana Q +r on ~ < camle Range (RA) = + X= CHART Sum pic Nor ——~—» UCL CUT-546) ct (4H) LCL C4e-d54)Ex: A Machine 1 set fe etiver the packels of a given werght: Ten samples of Aize five coene examined and tle follewing susulls wou obteined ; 43 49 37 44 45 37 5) 46 43 47 Calculate the contol Lincits foe X-chant & R-chant (dvaw) Comment on the stele of cenhol: (1=5, d,=2+3 26, 4,=0-064 Solution: gic 7 k=1d Ke LEK; = 1 (431491 --447) = 44-2 K to R= LsR= 1 (sierse.-..46)= 5+8 K 10 Ala 3-0 control Qimity ane given Fox X- chart fox R- chant Clo = K = UG.2 ClLcry= R= 58 Ullgy= ¥13R UCL egy= Rt 3G R = 12.263 %) ain cm a = U7+S4 - = W846 LCR py = R~ 3c R= 56-643 Lela: X- 3k a 4, fn =o (being ~vc) = 4o- 854 Commemt + K-Chant :- tla sample means counesponding fo sample 70-2. ane cuddide ef CL, the process average is oul'ef conhol Te suger: ELL pmesemer of assignable causes cf ecwalion which should be. factd and coaneeted: R-chant:. Nome of sample sarge (ies beyomd tle CL: of Releut tle procs veriability is far conkel é Simca eme of Gonbol Chest (x Charl ) shours Lack of conbel, the pacts cannot be eeeqaricted fra state of statistieed conhe|