200849116 ί九、發明說明: ; 【發明所屬之技術領域】 本發明是有關於一種製程統計分析系統,特別是有關 於一種具有查詢與篩選模組’可以對統計資料進行交又比 對”而生成製程統計分析結果之製程統計分析系統。 【先前技術】 為使生產線運作順暢,產品良率保持穩定,生產線必 須具有良好的製程統計系統,統計方法的應用、生產資料 的收集以及生產資料與統計管制條件的比對是製程統計系 、,不可缺少的-部份,另外製程統計⑽亦有修正各項製 程參數,將製程最佳化之功能,因此製程統計系統為提升 產〇口良率不可或缺的工具。 般生管人員需要不斷收集各項生產數據、機台狀 ^產扣良率等數據,再將各項進行各項統計分析,藉以 =忍生產線運作狀況以及產品良率是否保持於穩定的範 …、而以人貞如了統計分析常會有不穩定的狀況發生, Γ如對崎料進行蚊轉十分繁雜,常相為不同人 失對於各魏據的判斷標準不—,*造成統計分析之結果 3 ”不同人員對於丹常數據保留或剔除之判斷亦不相 同而異常數據容易對統計分析之結果具有相當大的影 5 200849116 :·響,另相人員進行各項統計分析,耗費大量時間,無法200849116 ί IX, invention description: [Technical field of invention] The present invention relates to a process statistical analysis system, in particular to a method with a query and screening module that can perform cross-matching of statistical data. Process statistical analysis system for process statistical analysis results [Prior Art] In order to make the production line operate smoothly and the product yield remains stable, the production line must have a good process statistics system, the application of statistical methods, the collection of production materials, and the production materials and statistical control. The conditional comparison is indispensable part of the process statistics system. In addition, the process statistics (10) also have the function of correcting various process parameters to optimize the process. Therefore, the process statistics system cannot improve the yield of the production pass. Lack of tools. The general management personnel need to continuously collect data such as production data, machine-like production yield, and other statistical analysis, so as to bear the operation of the production line and whether the product yield remains. Stable models... and people like statistical analysis often have unstable conditions, such as It is very complicated for Kawasaki to carry out mosquito reversal. The normal phase is different for people. The judgment standard for each Wei is not -, * results in statistical analysis. 3" Different people have different judgments on the retention or rejection of Dan Chang data and the abnormal data is easy. The results of the statistical analysis have a considerable impact. 5 200849116 :· Loud, other people carry out various statistical analysis, it takes a lot of time, can not
快速㈣統計分析結果,排除造成生產線運作不順暢Γ或 產品良率不穩定的原因。 S 例如在積體電路測試生產線中,生管人員於設計兮制 程時會預估-標準測試時間,當生產線穩定時,實際^ ,間會接近前述之標準測試時間,而當生產線不穩定時, 貫際測試_與標準測試時間產生較大的差異。而為 認測試狀況’生管人員收集測試數據,再依據生管人員的 涊知’將異常值剔除’以確認實際測試時間與標準測試時 :之間的差異。藉由生管人員進行上述分析,不但耗費大 里的人力以及時間,異常值的剔除容易因生管人員的認知 不同’而使分析結果失真,而且無法同時進行多批次的交 又比對,也不能得知機台與產品組合是否影響測試時間。 r餘^上述習知技術所存在的缺點,有必要提出一種製 :得到I析t統’可以對統計資料進行交又比對’並能快 的1/。#分析結果,藉以改善以人員進行各項統計分析 200849116 【發明内容】 本發明的目的在於提供—種製程統計分㈣統,可以 统計資料骑蚊比對,並能料彳㈣料分析結果, 藉以改善以人員進行各項統計分析的問題。 根據上述的目的’本發明揭露—種製程統計分析系 統’包含·複數之測試機台、—資料傳送系統以及一資料 庫为析系統。複數之測試機台,係用以對複數之積體電路 進行測試;該資料傳送系統連接至測試機台,用以傳送一 m 4貝料庫分析系統,連接至該資料傳送系統, 用以儲存和分析該賴㈣,其中,該:#料庫分析系統具 有一查詢與篩選模組’可輸人查詢條件與分析㈣值,對 該測試資料騎交又㈣,而生成縣斯分析結果。 朴本么月之製程統計分析系統,可以對統計資料: 灯又^比對’减快速得_計分析結果,㈣有效改 以人貝進仃各項統計分析的問題。根據本發明—實施 應用於積體電路_生產線,該製程統計分餘統可以1 統計貧料進行交又比對,錢快速得到統計分析結果,; 且可依據設定自動將㈣值剔除,不會產生因生管人員: 。不同@使分析結果失真的狀況,另外可同時 、又又比對’可以得知機台與產品組合㈣試時間€ 200849116 \影響。 【實施方式】 本發明的一些實施例將詳細描述如下。然而,除了如 下描述外,本發明還可以廣泛地在其他的實施例施行,且 本發明的範圍並不受實施例之限定,其以之後的專利範圍 為準。再者,為提供更清楚的描述及更易理解本發明,圖 式内各部分並沒有依照其相對尺寸繪圖,某些尺寸與其他 相關尺度相比已經被誇張;不相關之細節部分也未完全繪 出,以求圖式的簡潔。 第一圖顯示根據本發明一較佳實施例之製程統計分 析系統。該製程統計分析系統20包含:複數之測試機台 25、一資料傳送系統24、以及一資料庫分析系統21。 竣數之測試機台25,係用以對複數之積體電路進行測 試,並依據測試結果產生一測試資料,暫時儲存於該測試 機台25 ;資料傳送系統24,連接至該測試機台25,用以 傳送前述之測試資料;資料庫分析系統21,連接至該資料 傳送系統24,用以儲存和分析該測試資料,其中,該資料 庫分析系統21具有一查詢與篩選模組22,可輸入查詢條 件與分析篩選值,藉以對該測試資料進行交叉比對,並生 200849116 成一製程統計分析結果 圖。首先進行步驟I二資料庫分析系統的流程 32,輸入想要分析的口‘門程;接著進行步驟 開始時間與測試結束時/著t本J施例中為輸入測試 巧:述時間區間之測4批號資料了; J以=庫中 貧料庫中取得筛選* ·十气妾者進仃步驟34,由 5篩選值之測試資料,;算其平均驟L5二以前 37,計算過“ Vir二步驟 否結束,若尚。束39’確認計算迴圈是 進行步驟40,t$ 進仃34 i步驟38 ;最後 、、口束刀析 私,廷出製程統計分析結果。 在本實施例巾’係以積體電路測試生產線為例進行說 明,然而本發明並秘找此,亦朝於其他的各種生產 線、。積體料顺生產、㈣用崎完成製作的積體電路進 :測试’藉以確定其功能及品質,生管人員於設計該製程 k會預估-標準職時間,#生產線穩定時,實際測試時 間會接近前述之標準測試時間,而當生產線不穩定時,實 際測試時間與標準測試時間產生較大的差異。 第三圖顯示根據本發明一較佳實施例之製程統計分 析系統之-操作介面。查詢條件包含各項產品相關資料以 200849116 \及生產相關資料,例如產品相關資料可包含··客戶名稱、 產品型號等;生產相關資料可包含:測試開始時間、測試 結束時間、製程、機台群組、測試治具料號、測試治具編 唬以及測試程式等;分析篩選值可以包含生產相關設備資 料,例如機台群組、測試機台編號、測試治具料號、測試 程式以及批號等,藉由選取不同的分析篩選值,可以由萝 程統計分析結果得知生產相關設備以及其組合對測試時間 的影響,上述之分析篩選值可單選亦可複選,例如選取测 試機台編號可以得知測試機台編號對測試時間的影響,選 取測試治具料號與測試程式可以得知測試治具料號與測試 程式的組合對測試時間的影響。 另外該資料庫分析系統21具有—異常#料剔除選 擇’可將異常之該測試資料剔除後,再對該測試資料進行 父叉比對。當測試資料具有異常資料時,少數的異常資料 就可能會使統計分析的結果失真,無法代表其測試狀況, 因此必須對異常資料進行判別,並將異常資料剔除,根據 本實施例,上述之該異常資料剔除選擇,係以該測試資料 之標準差倍數進行選擇,例如丨倍標準差、2倍 望 例如選擇1倍鮮差時,資料庫分析㈣21會將大於平 均值加1倍標準差或小於平均值加i倍標準差的資料,認 200849116 m :·疋為異常資料並將其剔除,再進行統計分析。 第四圖顯不本發明一較佳實施例中,製程統計分析結 果之衣叙官制圖。該製程統計分析結果可包含一製程管制 ,縱軸爲製程測試數據,橫轴為時間,藉以使相關人員 容易判讀並了解製程測試數據之趨勢。 讀庫分析系統21可依據預先之設定,定時將該製 私統计分析結㈣電子郵件發駐相目人貞的電子信箱 28。根據本實施例’上述之#料庫分析系統η可依據—管 制上下限值,定時將超過該管制上下限值的製程統計分析 、、’σ果,以電子郵件寄^至相關人負的電子信箱。在本實 施例中,管制上下限值係設^為標準測試時間之倍數進^ 選擇’例如1_準測試時間、2〇%標準測試時間等。例如 選擇ίο%標準測試時間時,該資料庫分析系統2ι會將平均 測試時間切110%標準測試時間或小於9〇%標準測試時間 ,的製程統計分析絲,>xt子郵件Μ至相關人員的 信箱28〇 第五圖顯示本發明一齡存每> 孕乂仫貝鈿例中,以電子郵件發送 之製程統計分析結果。該製程統計分析結果包含各項產品 200849116 相關資料、生產相目資料以及測試統計資料,產品相 料可以包含:客戶名稱、產品型號等,·生產相關資料可以 包含.製程、機台群組、測試治具料號、測試程式;測試 統計資料可以包含:測試資料筆數、平均測試時間、平均 良率、測試時間最大值、測試時間最小值、測試時間標準 差、異常測試資料剔除數、剔除異常後平均時間、剔除異 常後標準差、鮮測試時間(標準工時)、以及差異率;其 中,差異率係設定為(平均測試時間_標準測試時間標 準測試時間。 ^ /利用本發明之製程、統計分析系統,可以對統計資料進 仃父又比對’並能快速得到統計分析結果,並且能自動將 超出管制上下限值的製程統計分析結果,以電子郵件通知 相關人員’可以有效改善以人員進行各項統計分析的問題。 上述之實施例僅係為說明本發明之技術思想及特點, 二的在使也悉此技蟄之人士能了解本發明之内容並據以 :加’當不能以之限定本發明之專利範圍,即凡其他未脫 離本發明所揭讀制完成之各種等效改㈣修飾都涵蓋 在本發明所揭露的範圍内,均應包含在下述之申請專利範 圍内。 200849116 【圖式簡單說明】 第一圖顯示根據本發明_ 析系統。 較佳實施例之製程統計分 圖中貧料庫分析系統的流程圖。 ‘明一較佳實施例之製程統計分 第二圖顯示根據第一圖中: 第三圖顯示根據本發明一 析系統之一操作介面。 第四圖顯示本發明一較佳實施例中,製程統計分析結 果之製程管制圖。 & 第五圖顯示本發明一較佳實施例中,以電子郵件發送 之製程統計分析結果。 【主要元件符號說明】 20 製程統計分析系統 21 資料庫分析系統 22 查詢與篩選模組 25 測試機台 24 資料傳送系統 28 電子信箱 31 開始 32 輪入時間區間 33 取得時間區間測試之批號資料 34 取得筛選值之測試資料 13 200849116 ·: 35 計算平均值與標準差 :36 過濾異常值 37 取得平均值與標準差 38 1己錄異常資訊 39 迴圈是否結束Quick (4) statistical analysis results, to rule out the reasons for the production line is not smooth, or the product yield is unstable. S For example, in the integrated circuit test production line, the production manager will estimate the standard test time when designing the process. When the production line is stable, the actual ^ will be close to the aforementioned standard test time, and when the production line is unstable, The continuous test _ is quite different from the standard test time. In order to confirm the test status, the health management personnel collect test data, and then remove the abnormal value according to the knowledge of the management personnel to confirm the difference between the actual test time and the standard test time. The above analysis is carried out by the management personnel, which not only consumes a lot of manpower and time, but also eliminates the abnormal value easily due to the different perceptions of the management personnel, and the analysis results are distorted, and it is impossible to carry out multi-batch cross-matching at the same time. It is not known whether the machine and product combination will affect the test time. r Yu ^ The shortcomings of the above-mentioned conventional techniques, it is necessary to propose a system: to get the I analysis of the statistic data can be compared and arbitrarily 'and can be faster 1 /. #分析结果。 The results of the analysis to improve the personnel to carry out various statistical analysis 200849116 [Description of the Invention] The object of the present invention is to provide a process statistics (four) system, statistical data can be compared to mosquitoes, and can be expected to analyze the results of four (four) materials, thereby Improve the issue of conducting statistical analysis by personnel. According to the above-mentioned object, the present invention discloses a process statistical analysis system comprising a plurality of test machines, a data transfer system, and a database as an analysis system. The plurality of test machines are used for testing a plurality of integrated circuits; the data transfer system is connected to the test machine for transmitting a m 4 bunker analysis system, connected to the data transfer system for storage And analysis of the Lai (four), wherein: #库库分析系统 has a query and screening module 'can enter the query conditions and analysis (four) values, the test data rides again (four), and generates county analysis results. The software statistical analysis system of Park Ben Mouyue can be used for statistical data: the light is compared with the 'reduction speed' and the results of the analysis are analyzed. (4) The problem of statistical analysis of the human beings is effectively changed. According to the present invention - the application is applied to the integrated circuit _ production line, the process statistic and the balance system can be compared and compared with the statistically poor material, and the money can quickly obtain the statistical analysis result; and the (four) value can be automatically removed according to the setting, and Produced by the birth control personnel: Different @ to make the analysis results distorted, in addition, can be simultaneously and in comparison 'can learn the machine and product combination (four) test time € 200849116 \ impact. [Embodiment] Some embodiments of the present invention will be described in detail below. However, the present invention may be widely practiced in other embodiments except as described below, and the scope of the present invention is not limited by the examples, which are subject to the scope of the following patents. Further, in order to provide a clearer description and a better understanding of the present invention, the various parts of the drawings are not drawn according to their relative dimensions, and some dimensions have been exaggerated compared to other related dimensions; the irrelevant details are not fully drawn. Out, in order to make the schema simple. The first figure shows a process statistical analysis system in accordance with a preferred embodiment of the present invention. The process statistical analysis system 20 includes a plurality of test machines 25, a data transfer system 24, and a database analysis system 21. The test machine 25 of the number is used to test the complex circuit of the plurality, and generate a test data according to the test result, temporarily stored in the test machine 25; the data transfer system 24 is connected to the test machine 25 For transmitting the foregoing test data; the database analysis system 21 is connected to the data transfer system 24 for storing and analyzing the test data, wherein the database analysis system 21 has a query and screening module 22 The query conditions are analyzed and the screening values are analyzed, so that the test data is cross-matched, and the results of the statistical analysis of 200849116 into one process are generated. First, the process 32 of the step I database analysis system is performed, and the port 'door process to be analyzed is input; then, the step start time and the end of the test are performed, and the input test is performed in the example of the test: the time interval is measured. 4 batch number information; J to = the library in the poor library to obtain the screening * · Ten gas entrants to enter the step 34, from the 5 screening value of the test data; calculate the average of the sudden L5 two before 37, calculated "vir The second step is not finished, if still. The bundle 39' confirms that the calculation of the loop is performed in step 40, t$进仃34 i step 38; finally, the mouth beam knife is analyzed, and the result of the statistical analysis of the process is obtained. 'The system is based on the integrated circuit test production line. However, the invention is also aimed at other various production lines. The integrated materials are produced, and (4) the finished circuit is finished with the test: By determining its function and quality, the management personnel will estimate the standard process time when the process is designed. When the production line is stable, the actual test time will be close to the standard test time mentioned above. When the production line is unstable, the actual test time is Standard test time production The third figure shows the operation interface of the process statistical analysis system according to a preferred embodiment of the present invention. The query conditions include various product related materials to 200849116 and production related materials, for example, product related data may include ··Customer name, product model, etc.; production related information may include: test start time, test end time, process, machine group, test fixture material number, test fixture compilation and test program; analysis and screening values can be Contains production-related equipment information, such as machine group, test machine number, test fixture number, test program, batch number, etc. By selecting different analysis and screening values, the production-related equipment can be known from the results of Luocheng statistical analysis. And the impact of the combination on the test time, the above analysis and screening values can be single or can be selected, for example, the test machine number can be selected to know the influence of the test machine number on the test time, select the test fixture number and test The program can know the effect of the combination of the test fixture number and the test program on the test time. The system 21 has an "abnormal # culling selection" to exclude the abnormal test data, and then the parent data is compared. When the test data has abnormal data, a small amount of abnormal data may cause statistical analysis. The result is distorted and cannot represent the test condition. Therefore, the abnormal data must be discriminated and the abnormal data must be eliminated. According to the embodiment, the abnormal data rejection selection is selected by using the standard deviation multiple of the test data, for example, When the standard deviation is 2 times, for example, when 1 time difference is selected, the database analysis (4) 21 will add more than 1 standard deviation or less than the average value plus i standard deviation of the average value, and recognize 200849116 m :·疋 as abnormal data. And it is excluded and statistical analysis is performed. The fourth figure shows the result of the statistical analysis of the process in the preferred embodiment of the invention. The statistical analysis results of the process can include a process control, the vertical axis is the process test data, and the horizontal axis is time, so that the relevant personnel can easily interpret and understand the trend of the process test data. The read library analysis system 21 can periodically send the electronic statistical analysis node (4) to the e-mail box 28 of the relevant person according to the preset setting. According to the present embodiment, the above-mentioned #库库分析系统η can be based on the control upper and lower limit values, and the timing statistical analysis of the process exceeding the upper and lower limits of the control, and the 'σ fruit, by e-mail to the electrons of the relevant person. mailbox. In the present embodiment, the control upper and lower limit values are set to a multiple of the standard test time, such as 1_quasi-test time, 2%% standard test time, and the like. For example, when ίο% standard test time is selected, the database analysis system 2 ι cuts the average test time by 110% standard test time or less than 9〇% of the standard test time, and the process statistical analysis wire, >xt sub-mail to related personnel The fifth box of the present invention shows the result of the statistical analysis of the process of sending an email by e-mail in the case of the invention. The statistical analysis results of the process include various products 200849116 related data, production related data and test statistics. The product materials can include: customer name, product model, etc. · Production related materials can include. Process, machine group, test Test item number, test program; test statistics can include: test data, average test time, average yield, test time maximum, test time minimum, test time standard deviation, abnormal test data rejection, rejection exception After the average time, after the abnormal standard deviation, fresh test time (standard working hours), and the difference rate; wherein the difference rate is set to (average test time _ standard test time standard test time. ^ / using the process of the present invention, The statistical analysis system can compare statistical data with the father and can quickly obtain the statistical analysis results, and can automatically report the results of the statistical analysis of the process that exceeds the upper and lower limits of the control, and notify the relevant personnel by e-mail to effectively improve the personnel. The problem of performing various statistical analyses. The above embodiments are only In order to explain the technical idea and the features of the present invention, the person skilled in the art can understand the contents of the present invention and according to the following: "When the patent scope of the present invention cannot be limited, that is, the other is not detached. The various equivalent modifications (4) modifications of the present invention are intended to be included in the scope of the present invention, and should be included in the scope of the following claims. 200849116 [Simple Description of the Drawings] The first figure shows according to the present invention. _ Analysis system. Flow chart of the poor database analysis system in the process statistics chart of the preferred embodiment. The second figure of the process statistics of the preferred embodiment is shown in the first figure: The invention is an operational interface of the system. The fourth figure shows a process control chart of the results of the statistical analysis of the process in a preferred embodiment of the present invention. <Fourth figure shows a preferred embodiment of the present invention, which is sent by e-mail. Process statistical analysis results. [Main component symbol description] 20 Process statistical analysis system 21 Database analysis system 22 Query and screening module 25 Test machine 24 Data transfer system 28 E-mail 31 Start 32 Round-in time interval 33 Batch number data for time interval test 34 Test data for screening value 13 200849116 ·: 35 Calculate the mean and standard deviation: 36 Filter outliers 37 Get the mean and standard deviation 38 1 Recorded abnormal information 39 Whether the loop is over