【正文】 rials Quality PCS Training Rev 3 May 21 2022 21 Observations Mean Moving Range S. D. Moving Range Subgroup 1 2 3 4 5 (X bar) (MRX ) (S) (MR S ) 1 2 3 : : 30 Average ? X Chart UCL (X) = X + = + () = CL (X) = X = LCL (X) = X = () = Example of Computing Control Limits for X S Chart ? S Chart UCL (S) = S + = + () = CL (S) = S = LCL (S) = S = () = 174。 Materials Quality PCS Training Rev 3 May 21 2022 25 Interpretation of X S Chart Some special causes of outofcontrol for ? X Chart – Changes in machine setting or adjustment – 參數(shù)設(shè)定被調(diào)整 – MStoMS technique inconsistent – Changes in material – 材料變化 ? S Chart – Machine in need of repair or adjustment – 機(jī)器須維修 – New Mses – Materials are not uniform – 材料一致性不夠 174。LCL (p) = p ? When LCL 0, put LCL = 0 or N/A ? Draw the control limits on p chart 174。 Materials Quality PCS Training Rev 3 May 21 2022 35 Caution When Using MR Method ? If there is autocorrelation, MR(Summary Stat) will underestimate the true process variation amp。 Materials Quality PCS Training Rev 3 May 21 2022 42 SPC Trend Rules Rule 4: 14 (or more) consecutive points are alternating up and down. 規(guī)則 4: 相鄰 14點(diǎn)規(guī)則性上下變動(dòng) Uses: Detects systematic effects, such as alternating machines, operators, suppliers, etc. 用途 : 察覺(jué)系統(tǒng)性的效性 . 如交替機(jī)臺(tái) ,人員 ,供應(yīng)商等 . Example: 174。增加成本 – They erode confidence in control charts as useful tools. Eventually all outofcontrol signals are ignored. As a result, quality, productivity, amp。 Materials Quality PCS Training Rev 3 May 21 2022 50 Treatment of OOC Points ? Every OOC points should have a pleted RFC where applicable amp。 Materials Quality PCS Training Rev 3 May 21 2022 Given example: UCLcurrent = LCLcurrent = Newly collected data resulted the following: UCLcalc = LCLcalc = ?runrun(calc) = MR / 3 = / 3 = UCL Change Ratio= (UCLcalc UCLcurrent) / ?runrun(calc) = LCL Change Ratio = (LCLcurrent LCLcalc) / ?runrun(calc) = = Indicates a need to change the current control limits! Change Ratio Example 174。 Materials Quality PCS Training Rev 3 May 21 2022 58 Control Limits ? Calculated from data, based on actual process performance ? Describe the natural range of performance of a stable process ? Describe the amount of natural process variation ? Used to determine process stability 174。 Materials Quality PCS Training Rev 3 May 21 2022 65 LSL USL A B Tail Closer to Spec Limit Cpk = B/A A = 3s B = LSL x ? Cpk pares the upper half of the distribution to the upper spec amp。 capability 174。 Process window characterization 187。 Materials Quality PCS Training Rev 3 May 21 2022 53 Change Ratio ? A change ratio may be utilized to detect when the current limits may be inappropriate 可以計(jì)算 change Ratio 來(lái)判斷目前管制界限是否合適 LCL Change Ratio = (LCLcurrent LCLcalc) / ?runrun(calc) UCL Change Ratio = (UCLcalc UCLcurrent) / ?runrun(calc) ? A large negative change ratio suggests tightening a control limit 若負(fù)值很大 ,建議縮小管制界限 ? A large positive change ratio suggests widening the control limit 若正值很大 ,建議放寬管制界限 Note: ?runrun(calc) = / 3 174。 Materials Quality PCS Training Rev 3 May 21 2022 49 Trend Rule Remendations ? Only use the trend rules that signal process instabilities for which you are capable of responding ? 在你有能力反應(yīng)處理的不穩(wěn)定制程 ,才使用趨勢(shì)規(guī)則 ? Justification needed for not using other SPC trend rules ? Std dev amp。 the cumulative false alarm rate can be very large ? 使用大量的判斷規(guī)則是不智的 ,因?yàn)槊恳粭l都有誤判的機(jī)率 ,累積起來(lái)將可能很大 ? False alarm rate is the frequency of control chart signals when nothing at all is actually wrong with the process ? 誤判機(jī)率指的是制程一點(diǎn)問(wèn)題都沒(méi)有 ,但規(guī)則確告訴我們有問(wèn)題 . ? False alarms are undesirable: ? 誤判警告並不是我們想要的 – They reduce productivity amp。 Materials Quality PCS Training Rev 3 May 21 2022 40 SPC Trend Rules Rule 2: 9 consecutive points on the same side of the centerline 規(guī)則 2: 連續(xù) 9點(diǎn)在中心線同側(cè) Uses: Detects small shifts or trends 用途 : 察覺(jué)微小 /趨勢(shì)性的偏差 False alarm rate: % 機(jī)率 :% Example: UCL CL LCL 174。 Materials Quality PCS Training Rev 3 May 21 2022 34 T i m eDataC o ns e c u t i v e l ot sa re v e ry s i mi l a rLo t s f a r a pa rt ma ybe v e ry dif f e re nt? Timerelated condition where consecutive data values are correlated (. dependent) 連貫的數(shù)據(jù)有關(guān)連性 ? Data values collected nearby in time are very similar 鄰近時(shí)間的數(shù)據(jù)非常相似 ? Data values collected far apart in time may be very different 較久時(shí)間的數(shù)據(jù)差異大 ? Tend to drift over time。UCL(p) = p + 187。 Materials Quality PCS Training Rev 3 May 21 2022 23 WARNINGS: Group / Summary will sort the new table in alphabetical order of the grouping variable. Control charts must always be plotted in time order. Therefore, if the summary table is not in time order, you will have to sort the table in correct time order before making the control chart. Example of Computing Control Limits for X S Chart using JMP LCL (S) = 0 174。CL (S) = S 187。顯示抽樣標(biāo)準(zhǔn)差的改變 174。 Materials Quality PCS Training Rev 3 May 21 2022 17 X S Chart Concept Consists of Two Portions: – X Chart 187。 Materials Quality PCS Training Rev 3 May 21 2022 13 Control Chart Classifications 類型 Classifications of control charts are depending on the type of data 依數(shù)據(jù)型態(tài)可分類為 : ? Variables data 計(jì)量型數(shù)據(jù) – A characteristic measured on a continuous scale resulting in a numerical value 特性 : 連續(xù)性 ,可量測(cè) ,有小數(shù)點(diǎn) – Examples: Void Size, Bond Pull Strength, Coplanarity, Ball Height, etc. 如尺寸 ,推拉力 ,平面度 ,高度 ? Attributes data 計(jì)數(shù)型數(shù)據(jù) – A characteristic measured by of conforming amp。 do not accurately indicate 3 sigma control limits. See your site statistician for advice on calculation methods when assumptions