TWI363257B - - Google Patents
Download PDFInfo
- Publication number
- TWI363257B TWI363257B TW095123119A TW95123119A TWI363257B TW I363257 B TWI363257 B TW I363257B TW 095123119 A TW095123119 A TW 095123119A TW 95123119 A TW95123119 A TW 95123119A TW I363257 B TWI363257 B TW I363257B
- Authority
- TW
- Taiwan
- Prior art keywords
- tester
- test
- semiconductor integrated
- integrated circuit
- mentioned
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/28—Testing of electronic circuits, e.g. by signal tracer
- G01R31/317—Testing of digital circuits
- G01R31/3181—Functional testing
- G01R31/3183—Generation of test inputs, e.g. test vectors, patterns or sequences
- G01R31/318314—Tools, e.g. program interfaces, test suite, test bench, simulation hardware, test compiler, test program languages
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/28—Testing of electronic circuits, e.g. by signal tracer
- G01R31/317—Testing of digital circuits
- G01R31/3181—Functional testing
- G01R31/319—Tester hardware, i.e. output processing circuits
- G01R31/31903—Tester hardware, i.e. output processing circuits tester configuration
- G01R31/31912—Tester/user interface
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Tests Of Electronic Circuits (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
- Test And Diagnosis Of Digital Computers (AREA)
Description
1363257 - 九、發明說明: 【發明所屬之技術領域】 • 本發明是關於一種伴隨半導體積體電路之開發的支援 系統。 【先前技術】 ,. 近年來’隨著半導體技術的發達,LSi (Large Scale φ intergratlon)等半導體積體電路的越來越辑雜和多樣,有 關其開發的效率也重要起來。 , 半導體積體電路的開,步驟亦即從設計到製造的步驟 一般如下。當開發泮導體積體電尨時.,首先,進行系統設 。十亦即以功此動作為中心的動作層次的設計。接著,進 行邏輯閘層次的邏輯設計和以元件層次表現該邏輯設計的 電路設計。然後,製作光罩,並轉移至稱為前製程的晶圓 製造製程,其利用該光罩在晶圓上形成半導體積體電路。 • 之後,以探針測試來試驗在前製程中所製造的晶圓· .,: · 判斷是否良好,將晶圓分割成晶片(半導體積體電路),篩 選出以探針測試出為良品的良品晶片’進行針對封裝的組 裝。 一. =組裝品藉由特性評估來偵錯,當其特性被辨認為滿 足既定的值,經由根據最後測試規格的測試程式所提供的 最後測試之後,轉移至量產。 在這二測試中,使用一種稱為測試器的裝置。測試器 為一台價值.高達數千萬曰圓至數億曰圓的昂貴裝置,使用 2036-8152-pp;Ahddub 5 iD/ 的問 成本也很高’測試器的選取對❹者而言是一個重要 題0 相:、·式器油測4頭、測試器本體、控制器等所構成。所 月:4式頭和作為測試對象的半導體積體電路的輸出入端 連接内建有複數個稱為引腳電子產品的介面板其中, 引腳電子產品對半導體積體電路輸出入訊號。 作為測S式器的架構,有複數個引腳共用時序產生電路 和圖案產生電路的共享資源方式、每個引腳具有時序產生 電路U圖案產生電路的逐?丨腳方式、每個引腳具有時 序產生電路和圖案產生電路的方式(以下稱為「全逐引腳方 式」)等。 過去’共享資源方式的測試器受到大量的使用,而最 近’逐引腳方式的測試器開始受到大量的錢。預測今後 全逐引聊方式的測試器會受到大量的使用。各種測試器具 有對半導體積體電路供給電力的電源、評估半導體積體電 路之輸出入端子之DC(Direct Current)特性的阢測量系 統、產生用來供給半導體積體電路之Dc準位的DAC(Digital1363257 - IX. Description of the Invention: [Technical Field of the Invention] The present invention relates to a support system accompanying the development of a semiconductor integrated circuit. [Prior Art], In recent years, with the development of semiconductor technology, semiconductor integrated circuits such as LSi (Large Scale φ intergratlon) have become more and more diverse, and the efficiency of their development is also important. The steps of the semiconductor integrated circuit, the steps from design to manufacture are generally as follows. When developing a volumetric body, first, the system is set up. Ten is the design of the action level centered on this action. Next, the logic design of the logic gate level and the circuit design of the logic design are performed at the component level. Then, a photomask is produced and transferred to a wafer fabrication process called a front process, which uses the photomask to form a semiconductor integrated circuit on the wafer. • Afterwards, test the wafers manufactured in the previous process with a probe test. • Determine whether the wafer is divided into wafers (semiconductor integrated circuits) and screen out the probes. Good wafers' are assembled for packaging. 1. The assembly is debugged by characterization, and when its characteristics are identified as meeting the established value, it is transferred to mass production after the final test provided by the test program according to the final test specification. In these two tests, a device called a tester was used. The tester is an expensive device with a value of tens of millions of rounds to hundreds of millions of rounds, using 2036-8152-pp; the cost of Ahddub 5 iD/ is also very high. 'The choice of tester is for the latter. An important question is 0 phase: , · 4 oil test, tester body, controller, etc. Month: The input and output terminals of the 4-type head and the semiconductor integrated circuit to be tested are connected to a plurality of dielectric panels called pin electronic products, in which the pin electronics output signals to the semiconductor integrated circuit. As the architecture of the S-type device, there are a plurality of pins sharing the shared resource mode of the timing generating circuit and the pattern generating circuit, and each pin has a timing generating circuit U pattern generating circuit. The method of squatting, each pin has a timing generating circuit and a pattern generating circuit (hereinafter referred to as "all-by-pin method"). In the past, the tester of the shared resource method was heavily used, and the recent 'pin-by-pin tester' began to receive a lot of money. It is predicted that the tester in the future will be heavily used. Each of the testers has a power supply for supplying power to the semiconductor integrated circuit, a measurement system for evaluating DC (Direct Current) characteristics of the input and output terminals of the semiconductor integrated circuit, and a DAC for generating a Dc level for supplying the semiconductor integrated circuit ( Digital
Analog Converter)轉換器等。 這些測5式器受到在CPU(Central Processing Unit)的 OSCOperating System)下動作的測試程式控制。此種測試 程式以所謂的測試器語言來撰寫》測試器語言一般隨著測 試器架構的不同而不同。作為此種測試器語言,當初出現 了為了直接控制硬體且稱為機器語言的組合語言。相對於 此’為了付到程式性尚的έ§· §,有人想出一種測試器控制 2036-8152-PF;Ahddub 6 1363257 語言,也開始使用FORTRAN形式和BASIC形式。再者,作為 結構性語言的pascal—時也受到廣泛的應用,不過,現在c 語言變成主流。 又,近年來,在半體導體積體電路的領域中,出現了 稱為無晶圓公司的電路設計公司、提供評價工具以在工作 站等電腦上評價電路設計公司所設計的半導體積體電路的 功能且被稱為EDA(自動設計工程)的公司、將與電路設計公 司所設計料導體積體電路有關測試程式轉換成可^行^ 試器的程式且被稱為測試屋的測試公司、根據電路設計公 司所設計的電路設計資料製作光罩的光罩製作公司、使^ 所製作的光罩製造(製作)半導體積體電路且被稱為晶圓代 工廠的晶圓製作公司、使用測試公司所製作的測試程式承 包所擁有的測試器的測試且被稱為測試工廠的公司等,這 些專業公司在半導體積體電路的開發方面有平面分工的: 勢。 因此,在專利文獻4,揭示一種關於商業模式的技 術,亦即,測試公司首先對EDA販賣商、電路設計公司及曰 圓製作公司提案測試器語言’接著,EDA販賣商對電路設: 公司提供進行半導體積體電路之邏輯模擬的程式,之後, 電路設計公司設計包括所要的邏輯功能的半導體積體電路 並檢驗虛擬測試器(在電腦上表現過測試器功能的測試器) 的邏輯功能,然後’透過網際網路將以測試器語言來製作 的測試程式傳送至測試公司,然後,測試公司根據該測試 程式選取可使用的測試器。 2036-8152-PF;Ahddub 7 1363257 [專利文獻1 ] 特開 2003-270305(段落 0033〜0035,第 6圖) 【發明内容】 【發明所欲解決的課題】 不過,在專利文獻1的技術中,在選取測試器之前的階 段’有許多專業公司的介入,所以,作業變得複雜或者沒 有效率,因此有花很多費用的問題。 因此’本發明赛於上述問題’目的在提供一種與開發 含有測試器之選取等的半導體積體電路有關且具有效率的 支援系統。 【用以解決課題的手段】 為了解決上述課題,本發明之半導體積體電路開發支 援系統包括使用測試器進行半導體積體電路之測試的受託 公司所擁有的控制裝置和域測試的委託者所擁有且上述 控制裝置透過網路來連接的終端裝置,上述控制裝置具有 記憶部和處理部’上述記憶部儲存指定並轉換上述測試所 需要的參數之後再將其使用於上述測試器中的測試器“ 程式、與複數個上述測試器有關且含有測試所需要的上述 參數的上述測試器之動作用資料的容許範圍被包括在内的 測試β資料,上述處理部根據來自 述終端裝置的請求, 將上述測試器語言程式傳送至上述終 , U. , 從上述終維 裝置接收根據上述測試器語言程式 1教作成的測試程式 6 2036-8152-PF;Ahddub 1363257 後’分析該測試程式的任意一部分並提取上述測試器的動 作用資料,藉由比較該動作用資料和儲存於上述記憶部的 上述測試器資料’提取可使用的測試器。 【發明效果】 本發明指出,可實現與包括測試器選取等的半導體體 積體電路的開發有關且有效率的支援系統。Analog Converter) converters, etc. These testers are controlled by a test program that operates under the OSCOperating System of the CPU (Central Processing Unit). Such test programs are written in a so-called tester language. The tester language generally varies with the tester architecture. As such a tester language, a combined language for direct control of hardware and called machine language appeared. In contrast to this, in order to pay for the procedural §§ §, someone came up with a tester control 2036-8152-PF; Ahddub 6 1363257 language, and also began to use FORTRAN form and BASIC form. Furthermore, pascal as a structural language has also been widely used, but now c language has become mainstream. Moreover, in recent years, in the field of a half-body volume circuit, a circuit design company called a fabless company has appeared, and an evaluation tool has been provided to evaluate a semiconductor integrated circuit designed by a circuit design company on a computer such as a workstation. A company that functions as EDA (Automatic Design Engineering), a test program that converts a test program related to a volumetric circuit designed by a circuit design company into a testable program, and is called a test house test company. The circuit design data designed by the circuit design company, the photomask manufacturing company that manufactures the mask, the photomask manufacturing circuit that manufactures the photomask, and is called the wafer fabrication company, the use test company. The testers that are produced by the test program contract test are called test factory companies, etc. These professional companies have a division of labor in the development of semiconductor integrated circuits: potential. Therefore, in Patent Document 4, a technique relating to a business model is disclosed, that is, a test company first proposes a tester language to an EDA vendor, a circuit design company, and a company, and then, an EDA vendor provides a circuit design: a program for performing a logic simulation of a semiconductor integrated circuit, after which the circuit design company designs a semiconductor integrated circuit including a desired logic function and verifies the logic function of the virtual tester (a tester that exhibits a tester function on a computer), and then 'Transfer the test program made in the tester language to the test company via the Internet. Then, the test company selects the tester that can be used according to the test program. 2036-8152-PF; Ahddub 7 1363257 [Patent Document 1] JP-A-2003-270305 (paragraphs 0033 to 0035, FIG. 6) [Explanation] [Problems to be Solved by the Invention] However, in the technique of Patent Document 1, In the stage before the selection of the tester, there are many professional companies involved, so the work becomes complicated or inefficient, so there are problems that cost a lot. Therefore, the object of the present invention is to provide a support system which is related to the development of a semiconductor integrated circuit including the selection of a tester and the like. [Means for Solving the Problem] In order to solve the above problems, the semiconductor integrated circuit development support system of the present invention includes a control device owned by a trust company and a domain test commissioner who is testing the semiconductor integrated circuit using a tester. And the terminal device connected to the control device via the network, wherein the control device includes a memory unit and a processing unit. The memory unit stores a parameter that specifies and converts parameters required for the test, and then uses the tester in the tester. a program, a test beta data included in an allowable range of the action data of the tester related to the plurality of testers and including the above-mentioned parameters required for the test, the processing unit according to the request from the terminal device The tester language program is transmitted to the above terminal, U., and receives the test program 6 2036-8152-PF created according to the tester language program 1 from the above-mentioned terminal device; Ahddub 1363257 and then 'analyzes any part of the test program and extracts The action data of the above tester, by comparing the action data The above tester data stored in the above-mentioned memory unit is used to extract a usable tester. [Effect of the Invention] The present invention indicates that an effective support system relating to the development of a semiconductor bulk circuit including a tester selection or the like can be realized.
【實施方式】 以下一面參照圖®來說日月本發明之半導體積體電路開 發支援系統。首先,參照第丨圖至第7圖來說明半導體積體 電路開發支援系統的結構。 1圖為半導體錢電路開發支援系、统的整體結構 導體積體電路開發支援系統在結構上包括測試 第 圖。半 公司(測試的受託公司Μ、電路設計公司(測試的委託公 司)2及晶圓製作公司3’這些公司所包括的各種裝置相互以 網際網路等網路4來連接。 此外,電路設計公司2和晶圓製作公司3分別只在圖中 顯不出一個,但亦可分別為複數個。 又,雖未特別圖示,在半導體積體電路的開發中 可和光罩製作公司、夾具製造公司等有關聯。 、 測試公司!在進行半導體積體電路㈣以時 裝置10’又,雖然未圖示, 枯徑制 ^ ,. ^ w , η ^ 隸有錢種機㈣測試器。 控制裝置U)為個人電腦等電腦裝置,具有用來和外部裝置 2036-8152-PF;Ahddub 9 • 進行通訊的通訊部11、由鍵盤等所構成且進行資料輸入的 輪入部12、由顯示部等所構成且進行資料輸出的輸出部 13、儲存各種資料的記憶部14、由cpu(Centrai[Embodiment] Hereinafter, the semiconductor integrated circuit development support system of the present invention will be described with reference to the drawings. First, the configuration of the semiconductor integrated circuit development support system will be described with reference to FIGS. 1 is a diagram showing the overall structure of the semiconductor money circuit development support system. The volumetric circuit development support system includes the test diagram. The semi-company (tested trustee company, circuit design company (testing company) 2 and wafer fabrication company 3') are connected to each other by a network such as the Internet. In addition, the circuit design company 2 and the wafer fabrication company 3 can only be shown in the figure, but they can also be plural. In addition, although not specifically illustrated, in the development of the semiconductor integrated circuit, it is possible to work with the mask manufacturing company and the fixture manufacturing company. The test company! In the semiconductor integrated circuit (4), the device 10', although not shown, the dry diameter system ^, . ^ w, η ^ is associated with the money machine (four) tester. A computer device such as a personal computer includes a communication unit 11 for communicating with an external device 2036-8152-PF, Ahddub 9 , a wheel-in portion 12 composed of a keyboard or the like and inputting data, and a display unit. The output unit 13 that outputs the data, the memory unit 14 that stores various materials, and the cpu (Centrai)
Unit)等所構成且進行各種運算處理的處理部π及作為處 理部15之運算區域的記憶體〗6。 電路設計公司2在進行半導體積體電路的電路設計 時,包括終端裝置20。終端裝置2〇為個人電腦等電腦裝置, 具有進行畫面顯示的顯示部21。 晶圓製作公司3製作作為半導體電路之基礎的晶圓,包 括亦為個人電腦等電腦裝置的終端裝置3〇。 第2圖為電路設計公司2的終端裝置2〇中的顯示部21的 畫面顯示例(請斟酌參照第丨圖)。在此,說明各個畫面(視 窗)的大致情況,有關細節,後面將會敘述。此外,在第2 圖中,同時顯示各個畫面(視窗),實際上以可分開來顯示。 又,即使在未特別標示出來的情況下,終端裝置2〇的通訊 φ 對象全部為控制裝置1 〇。 〜也視由22為終端裝置20的使用者輸入用來通訊的認 證用使用者名稱和密碼的畫面。 互動式編輯器視窗23為使用測試公司j所揭示的互動 式編輯器(用測試器語言製作的編輯器)、輸入半導體積體 電路的測试所需要的參數(施加電壓的值等)等並製作測試 程式的畫面。 在互動式編輯器方面’若使用C語言的函式語法,針對 藉由C語言來運作的測試器在測試器這邊事先準備c語言的 2〇36-8152-PF;Ahddub 10 1363257 函式定義’可在不用轉換的情況下在該測試器上被使用。 互動式編輯器儲存於控制裝置1〇的記憶部14中。 又’以互動式編輯器製作成的測試程式敘述了所需要 的參數,所以,對藉由c語言以外的語言來運作的測試器, 也有容易進行程式轉換的特徵。 於是’藉由使用此互動式編輯器’對於測試器的相關 知識知道得少的人來說,也可以製作出適當的測試程式。 測試圖案視窗24為電路設計公司2輸入用來檢驗半導 體積體電路之測試圖案的畫面。在此所謂測試圖案,是指 為了檢驗半導體積體電路而順著時間序列表現出模擬時之 各個參數之設定值的型態。 通中,在電路設計公司2,以一種藉由監控邏輯變化點 提取變化點的時序資訊和邏輯值並敘述測試圖案的時間驅 動方式,製作測試圖案❶不過,在設計階段,時間驅動方 式有用,而在此之前,則有無法讀取測試器的問題。因此, 在一定的週期間,需要分隔測試圖案,轉換為針對測試圖 案/步驟將測試圖案敘述為1/〇的資訊的碼率方式,使得測 試器變得可讀。 因此,在本申請案中,為電路設計公司2以時間驅動方 式輸入測試圖案並在測試公司丨轉換為碼率方式的結構。藉 此,可在測s式公司1組合碼率方式的測試圖案和後述之RTL 形是的半導體積體電路資料,在電腦上進行虛擬測試,並 透過網際網路等將該結果傳送至電路設計公司2,於是可實 現與半導體積體電路的開發有關且有效率的支援系統。 2036-8152-PF;Ahddub 1363257 • 此外,在以下的記載中,以時間驅動方式表示的測試 圖案有時僅稱為「圖案」,作為區別。 • 回到第2圖的說明’測試圖案視窗24為電路設計公司2 輪入圖案檔名 '選取圖案形式(文字、VCD(Value Change -Dump) 、 WGLCWaveform Generation Language)、 STIL(Standard Test Interface Language))、以時間驅動 方式輸入圖案的畫面。 導覽視窗2 5為根據從測試公司1的控制裝置丨〇傳送來 • 的資料顯示可使用之測試的一覽表、使用各個測試器時的 測試成本並藉由選取付款項目來選取實際使用的測試器的 畫面。 虛擬測試視窗2 6為在想進行虛擬測試(在電腦裝置上 所進行的虛擬测試)的情況下以RTL(Regis ter Trans is terA processing unit π configured to perform various kinds of arithmetic processing, such as a unit, and a memory unit 6 as a calculation area of the processing unit 15. The circuit design company 2 includes the terminal device 20 when performing circuit design of the semiconductor integrated circuit. The terminal device 2 is a computer device such as a personal computer, and has a display unit 21 that displays a screen. The wafer fabrication company 3 manufactures a wafer as a basis for a semiconductor circuit, and includes a terminal device that is also a computer device such as a personal computer. Fig. 2 is a screen display example of the display unit 21 in the terminal device 2 of the circuit design company 2 (please refer to the figure for reference). Here, the outline of each screen (view window) will be described, and details will be described later. Further, in Fig. 2, the respective screens (windows) are simultaneously displayed, and are actually displayed separately. Further, even if it is not specifically indicated, all of the communication φ objects of the terminal device 2 are the control device 1 〇. The screen for inputting the authentication user name and password for communication to the user of the terminal device 20 is also referred to as 22. The interactive editor window 23 is an interactive editor (an editor made in a tester language) disclosed by the test company j, a parameter (a value of an applied voltage, etc.) required for inputting a test of the semiconductor integrated circuit, and the like. Create a screen of the test program. In the interactive editor's use of C language function syntax, for the tester operated by C language in the tester side prepared c language 2〇36-8152-PF; Ahddub 10 1363257 function definition 'Can be used on this tester without conversion. The interactive editor is stored in the memory unit 14 of the control unit 1〇. Further, the test program created by the interactive editor describes the required parameters. Therefore, the tester operating in a language other than the C language has a feature that it is easy to perform program conversion. Thus, by using this interactive editor, it is also possible to create an appropriate test program for those who know less about the tester. The test pattern window 24 is a screen for the circuit design company 2 to input a test pattern for testing the semi-conductor volume circuit. Here, the test pattern refers to a pattern in which the set values of the respective parameters at the time of simulation are expressed in time series in order to examine the semiconductor integrated circuit. In the circuit design company 2, a test pattern is created by extracting the timing information and logic values of the change points by monitoring the logic change points and describing the time-driven manner of the test pattern. However, in the design stage, the time-driven mode is useful. Before that, there was a problem that the tester could not be read. Therefore, during a certain period of time, it is necessary to separate the test pattern and convert it into a rate mode in which the test pattern is described as 1/〇 of the test pattern, so that the tester becomes readable. Therefore, in the present application, the circuit design company 2 inputs the test pattern in a time-driven manner and converts it into a code rate mode at the test company. Thereby, the test pattern of the combined code rate method and the semiconductor integrated circuit data of the RTL described later can be tested on the computer, and the result can be transmitted to the circuit design through the Internet or the like. Company 2, therefore, can realize an efficient support system related to the development of semiconductor integrated circuits. 2036-8152-PF; Ahddub 1363257 • In the following description, the test pattern indicated by the time drive method may be simply referred to as a "pattern" as a difference. • Go back to the description in Figure 2 'Test Pattern Window 24 for Circuit Design Company 2 Wheeling Pattern Name' to select the pattern form (Text, VCD (Value Change - Dump), WGLCWaveform Generation Language), STIL (Standard Test Interface Language) ), input the pattern of the pattern in a time-driven manner. The navigation window 2 5 displays a list of tests that can be used based on the data transmitted from the control device of the test company 1, the test cost when using each tester, and selects the actual tester by selecting a payment item. Picture. The virtual test window 2 6 is in the case of a virtual test (virtual test performed on a computer device) in the case of RTL (Regis ter Trans is ter
Logic:直接型電晶體邏輯電路)方式來敘述半導體積體電 路時隨著該半導體積體電路的rTL檔名的輸入進行該指示 φ 的畫面。此外’在此,半導體積體電路以RTL方式為前提, 但在使用其他方式的情況下,會有配合該型是的畫面顯示。 設計視窗27為根據來自測試公司1的控制裝置丨〇的資 料接收顯示在半導體積體電路上採用DFT(Design forLogic: Direct type transistor logic circuit) A screen in which the indication φ is performed in accordance with the input of the rTL file name of the semiconductor integrated circuit when the semiconductor integrated circuit is described. Further, the semiconductor integrated circuit is premised on the RTL method. However, when other methods are used, there is a screen display in which the type is matched. The design window 27 is based on the data received from the control device of the test company 1 and displayed on the semiconductor integrated circuit using DFT (Design for
Testability :可測試性設計)時的半導體積體電路的面積 增加率並在想要DFT時伴隨RTL檔名的輸入進行該指示的畫 面。 共乘光罩視窗28為欲算出利用各個共乘光罩的測試成 本時進行該指示的畫面》 2036-8152-PF;Ahddub 12 1363257 不良邏輯閘映射視窗29為顯示晶圓上的不良部位(不 良邏輯閘的位置)的畫面。 第3圖為顯示各個測試器的規格的範例的表格。此表格 儲存於記憶部14(參照第1圖)中,用來表示最高動作頻率、 弓丨腳數等各個測試器的規格。 第4圖顯示各種方式之測試器共通的基本結構。如第4 圖所示,測試器300在構造上包括對作為測試對象的半導體 積體電路L供給電源電壓的電源單元3〇1、對半導體積體電 路L的輸入端子輸入訊號的驅動器3〇2、比較從半導體積體 電路L的輸出端子輸出的訊號和期望值訊號的比較器3〇3、 用來產生輸入至半導體積體電路1的測試資料及期望值的 圖案產生器304、用來產生輸入至帛導體積體電路L的訊號 的施加時序的時序產生器3〇5、藉由測試程式來進行各個控 制的控制器306、用來進行輸出引腳的電壓準位檢測等直流 測試的DC測試電路307。 此外,電源單元301及DC測試電路3〇7如同測試程式, 並非對測試器來說是不同的東西,又,測定半導體積體電 路L的技術不需要變更,可有共通的程式碼。 第5圖顯示測試器語言TL的範例。在此測試器語言 中,測試器程式碼50中的各個測試器資源(測試器的運作所 需要的參數被包括在内的測試器的動作用資料)和各個c語 言函式形式52對應。 例如’當測試器資源為元件電源51時,C語言函式形式 52為「VS {單元數,施加電壓,測定電流範圍籍位電流 2036-8152-Pp;Ahddub 13 1363257 上限’箱位電流下限}」。在此情況下,若要使用此測試器 語言TL輸入用測試器進行測試時所用的參數,單元數(電源 的單元編號)等各參數可輸入至其被顯示的位置。若未輸 入’半導體積體電路開發支援系統1〇〇〇所指定的預設參數 會被輸入’所以’製作測試程式的製作人員不需要理解所 有的參數。對於其他的測試器資源而言,情況亦相同。 第6(a)圖為模式圖,顯示rTL方式的半導體積體電路。 在半導體積體電路L中’於輸入端子68和輸出端子69之間, 以進行既定之邏輯動作的組合電路和複數個正反器 (FF)61交互組合出的組合電路線65(雖然圖示出一條,但其 他線路亦相同)來連接,又,所有的正反器61藉由時鐘線66 來連接。 藉由此結構,在半導體積體電路L中,和正反器61同 步’可以良好的順序處理從輸入端子68輸入的訊號,並從 輸出端子69輸出訊號。 第6(b)圖為模式圖,顯示在第6(a)圖所示之半導體積 體電路L中使用SCAN技術時對各個正反器61附加多工器 (MUX)62的狀態。 一個正反器61附加一個多工器62。多工器62由AND邏輯 閘621,622及OR邏輯閘623這三個邏輯閘所構成,測試電路 線63、SCAN電路線64及組合電路線65分別以如圖所示的方 式來連接。 藉由此種連接’在第6(a)圖的半導體積體電路L中,可 將各個正反器61鏈結在一起並使之連動,並可在測試時選 2036-8152-PF;Ahddub 14 1363257 出不良邏輯閘的位置》此外,此SCAN技術的細節記載在特 •開2003-1 49300號公報等文·獻中。 第7圖顯示光罩共乘晶片的結構。放大圖71放大了晶圓 70的一部分。在放大圖71中,晶片有10mmxl0mm(A晶片等)、 5nnnxl〇mm(B晶片、D晶片等)、5nnnx5miii(C晶片等)這三種。 近年來,半導體積體電路的製作發展出微型化製程, 使用複雜的裝置、多層的線路等,所以,變得非常昂貴。 於是’在同一晶圓上製作複數種半導體積體電路的共乘光 罩能有效降低半導體積體電路的製作成本。 關於半導體積體電路開發支援系統丨000,在以上完成 了結構的說明,接著,一邊參照第8圖至第丨7圖,一邊說明 其運作(請斟酌參照第1圖)。 第8圖為流程圖’顯示與半導體積體電路開發支援系統 1000有關的整體動作流程。 首先,測試公司1使用控制裝置丨〇公開與使用於半導體 • 積體電路之測試中的測試器之介紹等有關的網頁(HP)(步 驟S801)。 電路設計公司2使用終端裝置2〇製作與預定製作之半 導體積體電路有關的電路設計資料(步驟S8〇2),根據該電 晶圓’亦即’對晶圓Testability: The area of the semiconductor integrated circuit at the time of testability designation and the screen for the indication with the input of the RTL file name when DFT is desired. The common reticle window 28 is a screen for performing the instruction when calculating the test cost by using the respective commissure masks. 2036-8152-PF; Ahddub 12 1363257 The bad logic gate mapping window 29 is a defective portion on the display wafer (bad The screen of the position of the logic gate). Figure 3 is a table showing an example of the specifications of each tester. This table is stored in the memory unit 14 (see Fig. 1) and is used to indicate the specifications of each tester such as the highest operating frequency and the number of bows. Figure 4 shows the basic structure common to testers of various modes. As shown in FIG. 4, the tester 300 is configured to include a power supply unit 3 for supplying a power supply voltage to the semiconductor integrated circuit L as a test object, and a driver for inputting a signal to an input terminal of the semiconductor integrated circuit L. Comparing the signal and the expected value signal outputted from the output terminal of the semiconductor integrated circuit L to the comparator 3〇3, the pattern generator 304 for generating the test data and the desired value input to the semiconductor integrated circuit 1, for generating an input to The timing generator 3〇5 for guiding the timing of the signal of the volume circuit L, the controller 306 for performing each control by the test program, the DC test circuit for performing the DC test such as the voltage level detection of the output pin 307. Further, the power supply unit 301 and the DC test circuit 3〇7 are similar to the test program, and are not different for the tester. Moreover, the technique for measuring the semiconductor integrated circuit L does not need to be changed, and a common code can be used. Figure 5 shows an example of the tester language TL. In this tester language, each tester resource in the tester code 50 (the action data of the tester in which the parameters required for the operation of the tester are included) corresponds to each of the c language function forms 52. For example, when the tester resource is component power supply 51, the C language function form 52 is "VS {unit number, voltage applied, current range measurement current is 2036-8152-Pp; Ahddub 13 1363257 upper limit 'box current lower limit} "." In this case, to use the tester language TL to input the parameters used for testing with the tester, the parameters such as the number of cells (unit number of the power supply) can be input to the position where they are displayed. If the preset parameters specified in the semiconductor integrated circuit development support system are not input, the producer who created the test program does not need to understand all the parameters. The same is true for other tester resources. Fig. 6(a) is a schematic view showing a semiconductor integrated circuit of the rTL method. In the semiconductor integrated circuit L, a combination circuit 65 that performs a predetermined logical operation between the input terminal 68 and the output terminal 69 and a plurality of flip-flops (FF) 61 are alternately combined (although One is connected, but the other lines are the same) to connect, and all the flip-flops 61 are connected by the clock line 66. With this configuration, in the semiconductor integrated circuit L, the signal input from the input terminal 68 can be processed in a good order in synchronization with the flip-flop 61, and the signal is output from the output terminal 69. Fig. 6(b) is a schematic view showing a state in which the multiplexer (MUX) 62 is attached to each of the flip-flops 61 when the SCAN technique is used in the semiconductor integrated circuit L shown in Fig. 6(a). A flip-flop 61 is attached to a multiplexer 62. The multiplexer 62 is composed of three logic gates, an AND logic gate 621, 622 and an OR logic gate 623. The test circuit line 63, the SCAN circuit line 64, and the combined circuit line 65 are connected as shown in the figure. By means of such a connection 'in the semiconductor integrated circuit L of FIG. 6(a), the respective flip-flops 61 can be linked together and linked, and 2036-8152-PF can be selected during the test; Ahddub 14 1363257 The position of the bad logic gate. In addition, the details of this SCAN technology are described in the special article of the Japanese Patent Publication No. 2003-1 49300. Figure 7 shows the structure of the reticle shared wafer. A magnified view 71 magnifies a portion of the wafer 70. In the enlarged view 71, the wafer has three types of 10 mm x 10 mm (A wafer or the like), 5 nnn x 10 mm (B wafer, D wafer, etc.), 5 nnnx 5 miii (C wafer, etc.). In recent years, the fabrication of semiconductor integrated circuits has developed a miniaturization process, which uses complicated devices, multiple layers of wiring, etc., and therefore becomes very expensive. Thus, the fabrication of a plurality of semiconductor integrated circuits on the same wafer can effectively reduce the fabrication cost of the semiconductor integrated circuit. The semiconductor integrated circuit development support system 丨000 has been described above with reference to the above description, and the operation will be described with reference to Fig. 8 to Fig. 7 (please refer to Fig. 1 as appropriate). Fig. 8 is a flowchart showing the overall operation flow related to the semiconductor integrated circuit development support system 1000. First, the test company 1 uses the control device to disclose a web page (HP) relating to the introduction of the tester used in the test of the semiconductor integrated circuit (step S801). The circuit design company 2 uses the terminal device 2 to fabricate circuit design data relating to the predetermined semiconductor body circuit (step S8〇2), according to the wafer '
2036-8152-PF;Ahddub 路設計資料,委託晶圓製作公司3製作晶圓 製作公司3的終端裝置30傳送該主旨的資邻 1363257 又’亦可從控制裝置10下載互動式編輯器來使用。(步驟 S804)。此外,也可以不使用互動式編輯器而下載其他測試 程式製作工具。 又,進入測試公司1的網頁時需要認證,電路設計公司 2可透過顯示部21的認證視窗22(參照第2圖),輸入使用者 名稱和密碼’使用此環境。藉此,可提高資訊通訊的安全 等級。 接著’電路設計公司2使用終端裝置20—邊參照顯示部 21’ 一邊在互動式編輯器視窗23(參照第2圖)使用互動式編 輯器,一邊根據電路設計資料製作敘述半導體積體電路之 測試項目的測試程式,並以文本、VCD、WGL、STIL中任一 形式製作測試程式(步驟S805),將測試程式和圖案傳送至 測試公司1的控制裝置10(步驟S806)e在步驟別〇6中,使用 測試圖案視窗24(參照第2圖),圖案檔案的形式亦從4個之 中選取並傳送。 此外’使用互動式編輯器製作成的測試程式為根據電 子測量技術的敘述方式的程式碼,亦具有可從敘述内容輕 易辨識出敘述涵義的特徵。 在步驟S807中,測試公司1透過控制裝置丨〇的處理部 1 5 ’從接收自電路設計公司2的測試程式提取測試器資源 (有關細節’敘述於後面的第9圖)。又,透過控制裝置i 〇 的處理部15’從接收自電路設計公司2的圖案產生碼率方式 的測試圖案。然後,從預先儲存於控制裝置的硬體之中提 取硬體(以下將其稱為時序資源),該硬體可從該測試圖案 ^36-8152-PF;Ahddub 16 丄北3257 根據測域所要求的速度或規格產生時序訊號。再者,在圖 案產生器的使用硬體中提取測試所需要的各種圖案的長度 等。 當提取時序資源時,處理部15藉由分析測試圖案的全 部或任意一部分(例如10%),提取需要的資料。在與近年來 根據RTL同步設計的電路設計資料有關的測試圖案中藉由 分析其中的百分之幾到百分之幾十’可保證取得所需要的 資訊’所以’在此情況下可進行一部分的分析。 另外’當分析所有的測試圖案時,會花費許多時間和 費用’僅藉由一部分的分析,對測試公司1來說,相當省力。 接著’測試公司1使用控制裝置10搜尋(提取)可使用的 測試器’算出各測試器的測試成本(步驟S8〇8 :有關細節, 敘述於後面的第1〇圖)。 此外,在步驟S808中,若使用DFT,可提取出更便宜的 測試器’降低測試成本,同時,可算出使用DFT而導致的晶 片面積增加率,對電路設計公司2提示該資訊,有關該細 節,敘述於後面的第11圖。 又’在步驟S808中,有關使用共乘光罩之後的測試成 本的計算,敘述於後面的第12及第13圖》 接著’測試公司1使用控制裝置1 〇將可使用的測試器和 各測試器的測試成本傳送至電路設計公司2的終端裝置 20(步驟 S809 )。 接收該内容後,電路設計公司2透過終端裝置20將所接 收的可使用的測試器和各測試成本顯示於顯示部21 (參照 2036-8152-PF;Ahddub 17 1363257 第2圖的導覽視窗25)。 ψ 當電路設計公司2想執行虛擬測試時,使用終端裝置2〇 透過顯示部21的虛擬測試視窗26指定電路設計資料(RTL檔 案)名稱,在虛擬測試的項目中打勾,傳送至控制裝置 藉此’委託測試公司1執行虛擬測試(步驟S8丨〇 )。 測試公司1使用控制裝置1〇,根據測試程式、測試圖案 及電路設計資料,透過處理部15執行虛擬測試,將該結果 傳送至電路設計公司2的終端裝置20(步驟S8n)。 ^ 藉由執行此虛擬測試,可發現不適用於各個測試器的 測試圖案、時序等,接收該結果之後,電路設計公司2可適 當地修正測試程式、測試圖案等。 此外,當針對半導體積體電路的邏輯功能進行虛擬測 試時,該檢驗宜以碼率方式來進行。在此,所謂碼率方式, 是指與稱為時間驅動方式或時間事件方式之圖案產生方式 相反的圖案產生方式。具體來說,時間事件方式藉由監控 • 邏輯變化點提取變化點的時序資訊和邏輯值來產生圖=f 相對於此,碼率方式在一定的週期内分割圖案,將圖案步 驟變成1/0的資訊’以產生測試圖案。 又,在此,所使用的測試圖案也可以不是其全部的圖 案,可在可充分發現不良情況的範圍内,使用任意一部分 (百分之幾到百分之幾十),g,可將各個作業的時間縮 短,得到效率。 接著,電路設計公司2使用終端裝置20選取實際使用於 測試中的測試器,將其傳送至測試公的控制褒置1〇,藉 2036-8152-PF;Ahddub 18 1363257 • 此委託測試(步驟S812)。 然後,測試公司1根據需要,進行將測試程式轉換為測 試器的動作用程式、預備失具等測試作業的準備工作(步驟 S813)。 又,晶圓製作公司3接受來自電路設計公司2的有關晶 圓製作的委託(步驟S803),製作晶圓(步驟S8H),將該晶 圓移交至測試公司1 (步驟S81 5 )。 然後,測試公司1使用控制裝置10、測試器、測試程式 (根據需要轉換為動作用程式)、測試圖案 '電路設計資料、 夾具等,執行晶圓的測試(步驟S816),將晶圓和測試結果 移父至電路设计公司2 (步驟S817 )»測試結果可從控制裝置 10傳送至終端裝置2〇。 最後,在控制裝置10上,將測試器資源、測試結果等 各種資訊登錄於§己憶部14(步驟S818 )。測試公司1藉由活用 各種資訊,掌握與測試器的頻率、引腳數、圖案長度時 Φ 序資源等有關的電路設計公司2的需求動向,可適切地進行 與將來的測試器採購等有關的投資計畫、低成本的測試器 的開發等。 第9圖為流程圖,顯示在第8圖的步驟S8〇7中提取測試 器資源時的處理。 在控制裝置10上,處理部15輸入測試程式,亦即,將 接收自電路設計公司2的終端裝置2〇的測試程式(參照第2 圖的互動式編輯器視窗23)儲存於記憶部14(步驟S901)。 處理部15從測試程式的行數n = 〇開始(步驟S902),藉由 2036-8152-PF;Ahddub 19 1363257 n=n+1來搜尋第1行(步驟S903) ’判斷該行是否為pin程式 碼,亦即,是否為與引腳數有關的程式(步驟S9〇4)。 處理部15在該行為Pin程式碼時(在步驟S9〇4中為2036-8152-PF; Ahddub road design data, commissioned by the wafer fabrication company 3, and the terminal device 30 of the production company 3 transmits the title of the subject. 1363257 Further, an interactive editor can be downloaded from the control device 10 for use. (Step S804). Alternatively, you can download additional test authoring tools without using an interactive editor. Further, when the web page of the test company 1 is entered, authentication is required, and the circuit design company 2 can input the user name and password through the authentication window 22 (see Fig. 2) of the display unit 21 to use this environment. In this way, the security level of information communication can be improved. Then, the circuit design company 2 uses the terminal device 20 to refer to the display unit 21' while using the interactive editor in the interactive editor window 23 (see FIG. 2) to test the semiconductor integrated circuit based on the circuit design data. The test program of the project, and the test program is produced in any form of text, VCD, WGL, STIL (step S805), and the test program and pattern are transmitted to the control device 10 of the test company 1 (step S806) e at step 6 In the test pattern window 24 (refer to FIG. 2), the form of the pattern file is also selected and transmitted from the four. In addition, the test program created by the interactive editor is a code based on the description of the electronic measurement technology, and has a feature that can easily recognize the meaning of the narrative from the description. In step S807, the test company 1 extracts the tester resources from the test program received from the circuit design company 2 through the processing unit 15' of the control device (the details are described in the following ninth figure). Further, the processing unit 15' of the control device i 产生 generates a test pattern of a code rate method from the pattern received from the circuit design company 2. Then, the hardware is extracted from the hardware stored in advance in the control device (hereinafter referred to as a time series resource), and the hardware can be obtained from the test pattern ^36-8152-PF; Ahddub 16 丄北3257 according to the measurement domain The required speed or specification produces a timing signal. Furthermore, the lengths of various patterns required for the test, etc., are extracted from the used hardware of the pattern generator. When the timing resource is extracted, the processing section 15 extracts the necessary data by analyzing all or any part (e.g., 10%) of the test pattern. In the test pattern related to the circuit design data designed in synchronization with RTL in recent years, by analyzing a few percent to a few tens of percent, it is guaranteed to obtain the required information 'so' in this case. Analysis. In addition, when analyzing all the test patterns, it takes a lot of time and expense. By only a part of the analysis, it is quite labor-saving for the test company 1. Next, the test company 1 searches for (extracts) the usable tester using the control device 10 to calculate the test cost of each tester (step S8〇8: details, which will be described later in the first drawing). In addition, in step S808, if DFT is used, a cheaper tester can be extracted to reduce the test cost, and at the same time, the wafer area increase rate caused by using the DFT can be calculated, and the circuit design company 2 prompts the information about the detail. , described in the eleventh figure below. 'In step S808, the calculation of the test cost after using the common-shaping mask is described in the following 12th and 13th drawings." Next, the test company 1 uses the control device 1 and the tester and each test to be used. The test cost of the device is transmitted to the terminal device 20 of the circuit design company 2 (step S809). After receiving the content, the circuit design company 2 displays the received usable tester and each test cost on the display unit 21 via the terminal device 20 (refer to 2036-8152-PF; Ahddub 17 1363257, FIG. 2, the navigation window 25 ). ψ When the circuit design company 2 wants to perform the virtual test, the terminal device 2 specifies the circuit design data (RTL file) name through the virtual test window 26 of the display unit 21, tickes the virtual test item, and transmits it to the control device. This 'commissioned test company 1 performs a virtual test (step S8丨〇). The test company 1 uses the control device 1 to perform a virtual test by the processing unit 15 based on the test program, the test pattern, and the circuit design data, and transmits the result to the terminal device 20 of the circuit design company 2 (step S8n). ^ By performing this virtual test, test patterns, timings, and the like that are not applicable to each tester can be found. After receiving the result, the circuit design company 2 can appropriately correct the test program, test pattern, and the like. Furthermore, when performing a virtual test on the logic function of the semiconductor integrated circuit, the test should be performed in a code rate manner. Here, the code rate method refers to a pattern generation method that is opposite to a pattern generation method called a time drive method or a time event method. Specifically, the time event mode extracts the timing information and the logical value of the change point by the monitoring • logical change point to generate a graph = f. In contrast, the bit rate mode divides the pattern in a certain period, and the pattern step becomes 1/0. Information 'to generate test patterns. Further, here, the test pattern to be used may not be all of the patterns, and any part (a few percent to several tens of percent) may be used within a range in which a defect can be sufficiently found, and g may be used. The time for the work is shortened and the efficiency is obtained. Next, the circuit design company 2 uses the terminal device 20 to select the tester actually used in the test, and transmits it to the control device of the test unit, by 2036-8152-PF; Ahddub 18 1363257 • This commission test (step S812) ). Then, the test company 1 performs a preparation work for converting the test program into a test program, a preliminary test, and the like, as needed (step S813). Further, the wafer production company 3 accepts the request for the wafer production from the circuit design company 2 (step S803), creates the wafer (step S8H), and transfers the wafer to the test company 1 (step S81 5). Then, the test company 1 performs the wafer test (step S816) using the control device 10, the tester, the test program (converted into an action program as needed), the test pattern 'circuit design data, the jig, etc., the wafer and the test As a result, the parent is transferred to the circuit design company 2 (step S817). The test result can be transmitted from the control device 10 to the terminal device 2A. Finally, on the control device 10, various information such as tester resources, test results, and the like are registered in the memory unit 14 (step S818). The test company 1 grasps the demand trend of the circuit design company 2 related to the frequency of the tester, the number of pins, the length of the pattern, and the like, by using various kinds of information, and can appropriately carry out the purchase of the tester in the future. Investment plans, development of low-cost testers, etc. Fig. 9 is a flowchart showing the processing when the tester resource is extracted in step S8〇7 of Fig. 8. In the control device 10, the processing unit 15 inputs a test program, that is, the test program received from the terminal device 2 of the circuit design company 2 (see the interactive editor window 23 of Fig. 2) in the memory unit 14 ( Step S901). The processing unit 15 starts from the number of lines of the test program n = ( (step S902), searches for the first line by 2036-8152-PF; Ahddub 19 1363257 n=n+1 (step S903) 'determines whether the line is a pin The code, that is, whether it is a program related to the number of pins (step S9〇4). The processing unit 15 is in the behavior of the Pin code (in step S9〇4)
Yes) ’從引腳程式碼中提取引腳數的值,將其儲存於記憶 部14(步驟S905)。 處理部15在該行不為引腳程式碼時(在步驟別〇4中為 No),前進至步驟S906,判斷該行是否為…程式碼,亦即, 是否為與元件電源有關的程式。 鲁 處理部15在該行為VS程式碼時(在步驟Μοβ中為Yes) 'The value of the pin number is extracted from the pin code and stored in the memory unit 14 (step S905). When the line is not the pin code (No in step 4), the processing unit 15 proceeds to step S906 to determine whether the line is a ... code, that is, whether it is a program related to the component power supply. The processing unit 15 is in the behavior VS code (in the step Μοβ
Yes),從VS程式碼中提取電源數的值,將其儲存於記憶部 14(步驟 S907)。 此外,在此,測試器資源中只有引腳數和電源數被提 取數值並被儲存,但實際上亦可提取並儲存對其他的測試 器資源的值。 處理部15判斷該行是否為最後一行(步驟S9〇8),若不 φ 為最後一行(No),返回步驟S903進行反覆的處理,若為最 後一行(Yes)’結束處理。 如此,在控制裝置1 〇中,可從測試程式提取測試器資 源。 第1 〇圖為流程圖’顯示在第8圖的步驟S808中搜尋測試 器並算出成本時的處理。 在控制裝置1 0中’處理部丨5輸入測試器資源等,亦即, 將在第8圖的步驟S807所提取的測試器資源和時序資源健 存於記憶部14(步驟S1001)。 2036-8152-PF;Ahddub 20 ^63257 接著,處理部15從記憶部14取得測試器資訊,亦即, 取得與測試器的規格有關的資料(參照第3圖)(步驟 S1002)。 處理部15從測試器的機型n = 0開始(步驟sl〇〇3),藉由 n=nH來搜尋第!個測試器的規格(步驟Sl〇〇4),判斷測試器 資源的各個值是否滿足該測試器的測試器機型資源(測試 器資料)的條件(步驟Si〇〇5)。 處理部15在步驟S1005中,在滿足條件時(Yes),在輸 出部13顯示該測試器(可使用的測試器(步驟sl〇〇6),在不 滿足條件時(No),前進至步驟sl〇〇7。 處理部15在步驟S1 007中,判斷是否完成關於測試器的 所有機型的處理,若未完成(N〇),返回步驟S1004進行反覆 的處理,右元成(Yes) ’算出各個可使用的測試器的測試成 本並將其顯示於輸出部13(步驟sl〇〇8)。 當計算測試成本時,首先要求出測試所需時間。測試 斤需時間可考慮冑寫測言式程式時戶斤使用#冑#器語言的種 類來求,和戶斤使用的測試器的種類不太有相關性。測試 成本可藉由對所算㈣測試所需時間乘以各個測試器的單 位時間使用單價(3日圓/秒等)來求出。測試器的單位時間 使用單價處理部15可從記憶部14(參照第!圖)的表格中所 儲存的測試器的採購價格、使用時間、折舊期人事費等 來作適當的計算》 如此,可搜尋(選取)可使用的測試器,又可算出各個 測試器的測試成本。再者,除了測試成本之外,亦可算出 2036-8152-PF;Ahddub 21 1363257 並提示將使用的測試器的使用時間的估算費用β 第11圖為流程圖,顯示算出使用在第6(b)圖中所說明 之SCAN技術之DFT時的晶圓面積增加率時的處理。 首先,在控制裝置1 0中,處理部丨5輸入RTL,亦即,將 RTL形式的電路設計資料儲存於記憶部14 (步驟s丨丨〇丨)。 接著’處理部15進行邏輯合成,亦即,轉換為RTL形式 的電路設計資料的邏輯電路(步驟Sii〇2) » 接著’處理部15在邏輯電路中,計數包括正反器的總 邏輯閘數N(步驟S1103),又,計數其中的正反器數目n(步 驟S1104)。 處理部15藉由m=3xn計算在多工器(Μϋχ)中所使用的邏 輯閘數m(步驟S1105)。如第6(b)圖所說明,當使用%—技 術的DFT時,對一個正反器附加3個多工器,所以,可藉由 m=3xn求出邏輯閘數m。 最後’處理部15藉由m/N · 1〇〇算出DFT所導致的晶片面 積增加率U)(步驟S1106)。 如此,在控制裝置10上,可算出使用DFT時的晶片面積 增加率。 然後’若使用SCAN技術的DFT,隨著晶片面積的增加, 晶片製作的成本也跟著增加’不過,可減少測試所需要的 引腳數並降低測試頻率等,所以,可在第8圖的步驟S8〇8 中搜尋(提取)便宜的測試器來作為可使用的測試器,藉 此,可使測試成本變得便宜。 於是’電路設計公司2在第8圖的步驟S812中,即使對 2036-8152 -PF ; Ah.ddub 22 1363257 DFT的知識不充分,也可考慮晶片面積的增加所導致的晶片 製作成本的增加、測試成本的降低等各成本事項,適當地 判斷是否將使用DFT。此外,當電路設計公司2想使用DTF 時,可透過顯示部21的設計視窗27(參照第2圖),輸入DFT 的使用,將其傳送至測試公司1的控制裝置10。 第12圖為流程圖,顯示利用光罩共乘時的測試成本算 出處理。此外’以下所謂樣本數,是指以單位面積除以作 為對象的晶片之後的值,例如,在第7圖中,將單位面積設 為 25mm2(每邊 5mm) ’ A chip、B chip、C chip及 D chip的 樣本數分別為4,2,1及2 » 又’當利用在第12圖中所進行的光罩共乘時的測試成 本計算的處理與在第8圖的步驟S808中的測試成本計算相 對應,在此情況下,從第8圖的電路設計公司2到晶圓製作 公司3之間有關晶圓製作的委託(步驟S8〇3)在步驟S8〇8以 後的時點進行。 首先,電路設計公司2使用終端裝置2〇選取共乘光罩 (參照第2圖的共乘光罩視窗28),將該資料傳送至測試公司 1的控制裝置10。 接著,在測試公司1的控制裝置1〇中,處理部15將接收 自電路設計公司2的終端裝置2〇的共乘光罩(共乘光罩幻和 作為對象的晶片的樣本數輸入至記憶部14(步驟sl2〇l)。 接著,處理部15參照與儲存於記憶部丨4的共乘光罩有 關的資料庫,判斷在共乘光罩s中的樣本數是否從缺(步驟 S1202) 〇 2036-8152-PF;Ahddub 23 1363257 處理部15在共乘光罩中的樣本數不為從缺時(在步驟 S1202中為No)’結束處理,在共乘光罩中的樣本數為從缺 時(在步驟51202中為Yes),將該晶片登錄於與記憶部丨彳的 共乘光罩有關的資料庫中(步驟S1203)。藉由此登錄,該共 乘光罩S的空出面積會據此減少。 接著,處理部15算出整個共乘光罩s的測試成本(步驟 S1 204 :有關細節,敘述於後面的第丨3圖),該整個共乘光 罩S的測S式成本除以總樣本數,算出單位面積的測試成本, 藉由對該值乘以特定晶片的樣本數,算出該晶片的測試成 本(步驟S1 205)。 藉此’在控制裝置10上,可算出該晶片在共乘光罩s 中的測試成本。 第13圖為流程圖,顯示在第12圖的步驟S1204中的整個 光罩共乘光罩S的測試成本的運算處理。 首先,處理部15將在共乘光罩S的測試中所使用的機型 數Μ設為1(步驟S1301)。 接著’處理部15將共乘光罩產品編號亦即在共乘光罩S 上所製作的晶片的產品編號設為〇(步驟S1302),將總測試 時間Τ設為〇(步驟S1303) » 接著,處理部15藉由D = D+1對共乘光罩產品編號的值加 1 (步驟S1 304),參照記憶部14,判斷使用於該共乘光罩產 品編號的晶片的測試器是否已經登錄為使用於共乘光罩S 上的測試器(步驟S1305)。 若測試器已經登錄(在步驟S1305中為Yes),不需要變 2036-8152-PF;Ahddub 24 1363257 .更使用於共乘光罩s的測試器的機型數,又,若測試器尚未 登錄(在步驟S1305中為No),處理部15在記憶部14中將該測 試器登錄為使用於共乘光罩S的測試器(步驟S13〇6),使用 於共乘光罩S的測試器的機型數因此增加一個,所以,藉由 M = MH,更新μ的值(步驟S1307)。 接著’處理部1 5從記憶部1 4讀取每一樣本數的測試時 間t(步驟S1308),藉由T = T + tx樣本數(使用該測試器測試的 晶片的樣本數)更新T的值(步驟S1309)。 ® 接著,處理部15判斷該晶片是否為最後產品(步驟 S1310),若不為最後產品(N〇),返回步驟以3〇4進行反覆的 處理’若為最後產品(Yes) ’藉由TD=交換作業時間(丨次的 測試器交換所需要的作業時間)χ (M_丨)算出總作業交換時 間亦即交換測試器所需要的總時間(步驟s丨311)。 之後,處理部1 5藉由(T + TD ) X單價(單位時間的測試成 本)舁出整個共乘光罩S的測試成本(步驟S1312)。此外,若 φ 單價隨著各個測試器而不同,可算出整個共乘光罩S的測試 成本。 如此’在控制裝置10上,可算出在該光罩共乘中的整 個共乘光罩s的測試成本。 然後,使用同一共乘光罩内的各晶片的測試器的種類 (機型數)越少,整個共乘光罩S的測試成本越低,各晶片的 測試成本也和其成正比,跟著降低。於是,藉由第12圖及 第1 3圖的處理,當電路設計公司2要求更低測試成本的測試 器時,可使在同一共乘光罩内所使用的測試器的種類減 2036-8152-PF;Ahddub 25 1363257 少,測試公司1的作業也可以更有效率。 又’藉由在共乘光罩中進行測試,可提高晶片的良率 (良品的比率)。 第14圖為流程圖,顯示製作不良邏輯閘映射時的處 理。所謂不良邏輯閘映射’是指顯示在晶圓上的不良處(不 良邏輯閘的位置)的映射。在此所製作的不良邏輯閘映射作 為第8圖的步驟S817中的測試結果的一部分來使用。Yes), the value of the power source number is extracted from the VS code and stored in the memory unit 14 (step S907). In addition, here, only the pin count and the power supply number of the tester resource are extracted and stored, but in fact, the values of other tester resources can also be extracted and stored. The processing unit 15 determines whether or not the line is the last line (step S9〇8). If φ is the last line (No), the process returns to step S903 to perform the reverse processing, and if the line is the last line (Yes), the processing ends. Thus, in the control device 1, the tester resources can be extracted from the test program. The first diagram is a flowchart 'showing the processing when searching for the tester and calculating the cost in step S808 of Fig. 8. In the control unit 10, the processing unit 丨5 inputs a tester resource or the like, that is, the tester resource and the time-series resource extracted in step S807 of Fig. 8 are stored in the storage unit 14 (step S1001). 2036-8152-PF; Ahddub 20 ^ 63257 Next, the processing unit 15 acquires the tester information from the storage unit 14, that is, acquires the data relating to the specifications of the tester (see Fig. 3) (step S1002). The processing unit 15 starts from the model of the tester n = 0 (step sl1), and searches for the first by n=nH! The specifications of the testers (step S1〇〇4) determine whether the respective values of the tester resources satisfy the conditions of the tester model resource (tester data) of the tester (step Si〇〇5). In step S1005, when the condition is satisfied (Yes), the processing unit 15 displays the tester (a tester that can be used (step sl6)), and when the condition is not satisfied (No), the process proceeds to the step. Sl〇〇7. The processing unit 15 determines in step S1007 whether or not the processing of all the models of the tester is completed, and if it is not completed (N〇), returns to step S1004 to perform the reverse processing, and the right element is (Yes). Calculate the test cost of each tester that can be used and display it on the output unit 13 (step sl8). When calculating the test cost, first ask for the time required for the test. When the program is used, the type of the language used by the #胄# device is not relevant to the type of tester used by the household. The test cost can be multiplied by the time required for the test (4) multiplied by each tester. The unit time is obtained by using the unit price (3 yen/sec, etc.). The unit price of the tester can be used by the unit price processing unit 15 from the table of the memory unit 14 (see Fig.). Depreciation staff costs, etc. Do the appropriate calculations. In this way, you can search (select) the testers that can be used, and calculate the test cost of each tester. In addition, in addition to the test cost, you can also calculate 2036-8152-PF; Ahddub 21 1363257 and Prompt for estimating the usage time of the tester to be used. Fig. 11 is a flowchart showing the processing when calculating the wafer area increase rate when using the DFT of the SCAN technique described in Fig. 6(b). In the control device 10, the processing unit 丨5 inputs the RTL, that is, stores the circuit design data in the RTL format in the memory unit 14 (step s丨丨〇丨). Then the processing unit 15 performs logic synthesis, that is, a logic circuit for converting circuit design data in the form of RTL (step Sii 〇 2) » Next, the processing unit 15 counts the total number of logic gates N including the flip-flops in the logic circuit (step S1103), and counts the same The number of flip-flops n (step S1104) The processing unit 15 calculates the number of logical gates m used in the multiplexer by m = 3xn (step S1105). As illustrated in Fig. 6(b), Attach a flip-flop when using the %-technical DFT Since three multiplexers are used, the logical gate number m can be obtained by m = 3xn. Finally, the processing unit 15 calculates the wafer area increase rate U) caused by the DFT by m/N · 1 ( (step S1106). . As described above, the control device 10 can calculate the wafer area increase rate when the DFT is used. Then, if the DFT of SCAN technology is used, the cost of wafer fabrication will increase as the area of the wafer increases. However, the number of pins required for testing can be reduced and the test frequency can be reduced. Therefore, the steps in Figure 8 can be used. In S8〇8, a cheap tester is searched (extracted) as a usable tester, thereby making the test cost cheap. Therefore, in the step S812 of Fig. 8, even if the knowledge of 2036-8152-PF; Ah.ddub 22 1363257 DFT is insufficient, the increase in the wafer fabrication cost due to the increase in the wafer area can be considered. Each cost item such as a reduction in test cost is appropriately judged whether or not DFT will be used. Further, when the circuit design company 2 wants to use the DTF, the use of the DFT can be input through the design window 27 (see FIG. 2) of the display unit 21, and transmitted to the control device 10 of the test company 1. Fig. 12 is a flow chart showing the test cost calculation processing when the reticle is used for sharing. In addition, the number of samples referred to below is a value obtained by dividing the unit area by the target wafer. For example, in Fig. 7, the unit area is 25 mm 2 (5 mm per side) ' A chip, B chip, C chip And the number of samples of the D chip are 4, 2, 1 and 2 respectively » and 'the processing of the test cost calculation when using the reticle sharing performed in Fig. 12 and the test in the step S808 of Fig. 8 Corresponding to the cost calculation, in this case, the request for wafer fabrication from the circuit design company 2 of FIG. 8 to the wafer fabrication company 3 (step S8〇3) is performed at a time point after step S8〇8. First, the circuit design company 2 selects the shared photomask (refer to the sharing mask window 28 of Fig. 2) using the terminal device 2, and transmits the data to the control device 10 of the testing company 1. Next, in the control device 1 of the test company 1, the processing unit 15 inputs the number of samples of the shared photomask (the shared photomask phantom) received from the circuit design company 2 into the memory. The processing unit 15 determines whether or not the number of samples in the common-sharing mask s is absent (step S1202) by referring to the database relating to the sharing mask stored in the memory unit 4 (step S1202). 〇2036-8152-PF; Ahddub 23 1363257 The processing unit 15 ends the processing when the number of samples in the common-sharing mask is not missing (No in step S1202), and the number of samples in the common-sharing mask is from When there is no time (Yes in step 51202), the wafer is registered in a database related to the sharing mask of the memory unit (step S1203). By registering, the sharing mask S is vacated. The area is reduced accordingly. Next, the processing unit 15 calculates the test cost of the entire commissure mask s (step S1 204: details, which will be described later in FIG. 3), and the S-type of the entire commissure mask S The cost divided by the total number of samples, the test cost per unit area is calculated by multiplying the value by the specific The number of samples of the slice is used to calculate the test cost of the wafer (step S1 205). By this, the test cost of the wafer in the shared photomask s can be calculated on the control device 10. Fig. 13 is a flow chart showing The calculation processing of the test cost of the entire photomask sharing mask S in step S1204 of Fig. 12. First, the processing unit 15 sets the number of models used in the test of the common-sharing mask S to 1 (step S1301) Next, the processing unit 15 sets the product number of the wafer produced on the common-sharing mask S, that is, the shared mask product number to 〇 (step S1302), and sets the total test time Τ to 〇 (step S1303). Then, the processing unit 15 adds 1 to the value of the shared mask product number by D = D+1 (step S1 304), and refers to the memory unit 14 to determine the test of the wafer used for the shared mask product number. Whether the device has been logged in as the tester used on the common reticle S (step S1305). If the tester has logged in (Yes in step S1305), it is not necessary to change 2036-8152-PF; Ahddub 24 1363257. The number of models of the tester shared by the reticle s, and, if the tester is not logged in (in step In No. S1305, the processing unit 15 registers the tester in the memory unit 14 as a tester used in the sharing mask S (step S13〇6), and is used in a tester of the shared mask S. Therefore, the number is incremented by one, so the value of μ is updated by M = MH (step S1307). Next, the processing unit 15 reads the test time t of each sample number from the memory unit 14 (step S1308), by The T = T + tx sample number (the number of samples of the wafer tested using the tester) updates the value of T (step S1309). ® Next, the processing unit 15 determines whether the wafer is the last product (step S1310), and if it is not the last product (N〇), the returning step repeats the processing with 3〇4 'if the last product (Yes)' by TD = exchange work time (time required for the exchange of testers) χ (M_丨) Calculate the total job exchange time, that is, the total time required to exchange the testers (step s 311). Thereafter, the processing unit 15 extracts the test cost of the entire commissure mask S by the (T + TD ) X unit price (test cost per unit time) (step S1312). In addition, if the unit price of φ varies with each tester, the test cost of the entire shared mask S can be calculated. Thus, on the control device 10, the test cost of the entire commissure mask s in the reticle sharing can be calculated. Then, the fewer the types of testers (the number of models) using the wafers in the same sharing mask, the lower the test cost of the entire shared mask S, and the test cost of each wafer is proportional to it, followed by lowering . Thus, with the processing of Figures 12 and 13, when the circuit design company 2 requires a lower test cost tester, the type of tester used in the same shared mask can be reduced by 2036-8152. -PF; Ahddub 25 1363257 Less, the test company 1's work can also be more efficient. Moreover, by performing tests in a shared photomask, the yield of the wafer (the ratio of good products) can be improved. Figure 14 is a flow chart showing the processing when creating a bad logic gate map. The term "bad logic gate map" refers to a map showing defects on the wafer (the position of the defective logic gate). The bad logic gate map created here is used as part of the test result in step S817 of Fig. 8.
控制裝置10的處理部15輸入得自於在第8圖的步驟 S816中所實施的測試的不良資料(不良圖案),亦即,將其 儲存於記憶部14中(步驟S1401)。 接者,處理部15參照與儲存於記憶部14的晶圓上的發 輯閘有關的資料庫亦即邏輯元件參考手冊(步驟si4〇2),右 定不良邏輯間(步驟S14〇3)e此外,在邏輯元件參考手* 中,不良資料和不良邏輯閘相互關連。 接著處理αΡ 1 5取得儲存於記憶部1 4的晶圓的佈局言 計資訊(步驟S14G4)’^不良邏輯閘的位置(步驟S1405) 然後,處理部15判斷盆曰太*田 斯,、疋否為最後一個不良資料,t 即’是否完成了所有的不良咨姓从占 小艮貝枓的處理(步驟S14〇6),若习 為最後一個不良資料. ’返回步驟14〇1進行反覆的肩 理,若為最後一個不良眘姐〜 、 育枓(Yes),結束處理。 如此,製作出不良玀赭 ^私閘映射’在第8圖的步驟S81 中,測試公司1的控制裝置】n 置10將該不良邏輯閘映射傳送至駕 路設計公司2的終端裝置2 在終端裝置20的顯示部21上| 示出該不良邏輯閘映射(參昭 麥…第2圖的不良邏輯閘映射視甚 2036-8152-PF;Ahddub 1363257 29)。此外,此不良邏輯閘映射可製作在各個晶片上,也可 製作在整個晶圓上。 然後,電路設計公司2可藉由檢視此不良邏輯閉映射, 推測或決定在各個晶片的製程中是否產生自某種原因(如 異物混入等)’藉由將該原因傳送至晶圓製作公司3,可對 其進订更有效率的改善,並提高半導體積體電路製作中的 良率(良品的比率)。 接著,一邊參照第15圖,一邊說明導覽視窗2 5 (參照第 2圖)的變形例(請斟酌參照第1圖)。在第15圖中,(a)顯示 校正資源參數前的導覽視窗25a,(b)顯示校正該參數後的 導覽視窗25a。 在第10圖的步驟S1 006中,控制裝置ι〇(參照第1圖)的 處理部15(參照第1圖)如同導覽視窗25(參照第2圖)對輸出 部1 3 (參照第1圖)顯示可使用的測試器,不過,如第15 ( a ) 圖所示,亦不這樣顯示,而如同導覽視窗25a那般顯示測試 器一覽表。 在第15(a)圖中,測試器a, B,C和其是否可被使用一 起被顯示出來。在此,當操作者選取不可使用(X :亦可以 其他符號顯示)的測試器B,在其下的攔位中,針對各個資 源(原因)以NG或OK來顯示與測試器B的參數設定是否恰當 有關的判斷結果。 在此’資源中只有頻率被判斷為NG,藉由校正該頻率 的參數,如第15(b)圖所示,頻率的判斷結果變為OK,於是, 測試器B變成可使用狀態。 2036-8152-PF;Ahddub 27 1363257 此外,當頻率的參數校正在參數的限值在50奈秒以上 的情況下,若設定為40奈秒而為NG狀態,藉由將該值改寫 為60奈秒等,可校正過來。 如此,藉由導覽視窗25a之類的顯示,操作者可輕易地 知道測試器在不可使用狀態下的原因,藉由改寫參數等輕 易使該測試器變成可使用狀態。 此外,此種導覽視窗25a的顯示、參數的改寫等在第8 圖的步驟S809中’可在顯示可使用的測試器等的電路設計 公司2的終端裝置20的顯示部21上進行。 接著,一邊參照第16圖,一邊說明在測試器上所使用 的指令的轉換(請斟酌參照第1圖)。第16圖為說明圖,顯示 在測試器上所使用的指令的轉換。 在第8圖的步驟S816中,測試公司1使用測試器等執行 晶圓的測試,不過此時,在測試器上所使用的指令可以不 是第16圖中的電路設計公司2的終端裝置2〇的顯示部。上 所表示的「VS」等指令(參照第5圖),而是預先轉換並儲存 於第16圖中的測試公司i的控制裝置1〇的記憶部14的 「GTL-VS」等指令。 換言之,在電路設計公司2的終端裝置20的顯示部21 上,顯示「VS」等指令(參照第2圖的互動式編輯器視窗⑶, 在測試公司i的控制裝置10上,當處理部15將所有這些指令 儲存於記憶部14時,可附w個字元以上(在此^個字元) 的字首或字尾(在此為「GTL-」)。 然後’在測試器這一端,斜试士 Λ ^對所有的指令附加所謂的字 2036-8152-PF;Ahddub 1363257 首或字尾,事先解讀並執行這些指令。 如此,即使「vs」、「ST〇p」等指令為在某種程式化 ^中代表其他涵義的保留字(預先決定的指令)等,在測 試器這H以毫無障礙地解讀這些指令,執行測試程式。 又’右字首或字尾的字元可迴避保留字的内容,可自 由組合英數字元、連字號(_)等來決定。 接著,邊參照第17圓,一邊說明在測試器上所使用 的函式定義的選取(請斟酌參照第!圖等)。第17圖為流程 圖,顯示與函式定義的選取有關的處理。 例如,如前所述,在第8圖的步驟S813令,測試公司工 進行測試器的測試作業的準備,此時,從控制裝置ι〇到測 S式器’可僅儲存測試所需要的函式定義。 如第17圖所示,首先,控制裝置10的處理部15從記憶 部14讀取接下來將使用的預定的測試程式(步驟si7〇i)。 處理部15設定i = l(步驟sl7〇3),讀取第“于的敘述(程 式瑪)(步驟S1705),判斷該敘述所需要的函式定義是否在 選取函式定義檔案中(步驟S1 707)e此外,所謂選取函式定 義槽案’疋指此時的測試程式的執行所需要的函式定義被. 收集在一起的檔案,其儲存於記憶部14中。 若需要的函式定義不在選取函式定義檔案中(在步驟 S1707中為No) ’處理部15從記憶部14的函式定義庫提取該 函式定義’將其登錄於選取函式定義檔案中(步驟sl7〇9)。 此外,所謂函式定義庫’是指預先使用且具有可能性的許 多函式定義被收集在一起的資料庫。 2036-8152-PF;Ahddub 29 1363257 • 若需要的函式定義在選取函式定義檔案中(在步驟 S1707中為Yes)而且在步驟sl7〇9之後,處理部15判斷該敘 述是否為最後一行(步驟S1 711)。 若該敘述不為最後一行(在步驟31了11中為No),處理部 15設定i = i + i(步驟S1713),返回步驟sl7〇5。 若該敘述為最後一行(在步驟81711中為Yes),處理部 1 5結束處理。 如此’控制裝置10從許多函式定義被收集在一起的函 籲式定義庫(例如函式定義數為1 000),製作僅收集此時測試 程式的執行所需要的函式定義(例如,函式定義數為2〇〇) 的選取函式定義庫,然後,將其傳送至測試器。 藉此’測試器可僅儲存需要的函式定義而不儲存不需 要的函式定義,所以,可減少額外負擔等,並縮短測試所 需要的處理時間。 此外’在此’需要的函式定義的選取(提取)及使用在 φ 使用測試器進行測試時進行,不過,除此之外,亦可.在進 行前述虛擬測試(第8圖的步驟S811)時進行。 以上元成了實施型態的說明,不過’本發明的型態並 不限定於此。 例如’本發明不僅可應用於LSI等半導體積體電路,亦 可應用於記憶體等使用半導體技術的所有裝置。 又’在DFT方面,不僅可使用SCAN技術,亦可使用 BIST(Built-In Self-Test)等其他技術。 再者’當各裝置透過網路4收發各種資料時,不僅可藉 2036-8152-PF;Ahddub 30 1363257 由密碼等技術,亦可藉由加密等技術來來提高安全等級。 除此之外,有關硬體、流程圖等具體結構,在不脫離 本發明之主要精神的範圍内,可作適宜變更。 【圖式簡單說明】 第1圖為半導體積體電路開發支援系統的整體結構圖。 第2圖為電路設計公司2的終端裝置2〇中的顯示部以的 畫面顯示例。 第3圖為顯示各個測試器的規格的範例的表格。 第4圖顯示各種方式之測試器共通的基本結構。 第5圖顯示測試器語言TL的範例。 第6(a)圖為模式圖,顯示RTL方式的半導體積體電路。 第6(b)圖為模式圖,顯示在第6(a)圖所示之半導體積體電 路L中使用SCAN技術時對各個正反器61附加多工器(Μυχ)62 的狀態。 第7圖顯示光罩共乘晶片的結構。 第8圖為流程圖,顯示與半導體積體電路開發支援系統 1 000有關的整體動作流程。 第9圖為流程圖,顯示提取測試器資源時的處理。 第10圖為流程圖,顯示在第8圖的步驟S8〇8中搜尋測試 器並算出成本時的處理。 第11圖為流程圖.,顯示算出使用SCAN技術之DFT時的晶 圓面積增加率時的處理。 第12圖為流程圖,顯示利用光罩共乘時的測試成本算 31 -PF;Ahddub 1363257 出處理。 第13圖為流程圖’顯示在第12圖的步驟sl2〇4中的整個 光罩共乘光罩S的測試成本的運算處理。 第14圖為流程圖,顯示製作不良邏輯閘映射時的處理。 第15(a)圖顯示校正參數前的導覽視窗253,第 圖顯示校正參數後的導覽視窗253。 第1 6圖為說明圖’顯示在測試器上所使用的指令的轉 換。 第17圖為流程圖,顯示與函式定義的選取有關的處理。 【主要元件符號說明】 1〜測試公司; 2〜電路設計公司; 3〜晶圓製作公司; 4〜網路; 10〜控制裝置; 11〜通訊部; 12〜輸入部; 13〜輸出部; 14〜記憶部; 15〜處理部; 16〜記憶體; 20〜終端裝置; 21〜顯示部; 22〜認證視窗; 2 5〜導覽視窗; 23〜互動式編輯器視窗; 25a〜導覽視窗; 2 4〜測試圖案視窗; 26〜虛擬測試視窗; 27〜設計視窗; 28〜共乘光罩視窗; 29〜不良邏輯閘映射視窗. 3 0〜終端裝置; 50〜測試器程式碼; 51〜元件電源; 52〜C語言函式形式; 2036-8152-PF;Ahddub 32 1363257The processing unit 15 of the control device 10 inputs the defective material (defective pattern) obtained from the test executed in step S816 of Fig. 8, that is, it is stored in the storage unit 14 (step S1401). The processing unit 15 refers to the logical element reference manual (step si4〇2), which is a database related to the issue gate stored on the wafer of the storage unit 14, and sets the bad logic between right (step S14〇3). In addition, in the logic component reference hand*, bad data and bad logic gates are related to each other. Then, the processing unit 15 obtains the layout information of the wafer stored in the memory unit 14 (step S14G4) '^ the position of the defective logic gate (step S1405). Then, the processing unit 15 determines the basin 曰太*田斯, 疋Whether it is the last bad information, t is 'whether it has completed all the bad consultations from the treatment of Xiao Biaobei (step S14〇6), if it is the last bad data. 'Return to step 14〇1 to repeat Shoulder, if it is the last bad sister ~, Yu Yu (Yes), the end of the treatment. In this way, a defective 私^ private gate map is created. In step S81 of Fig. 8, the control device of the test company 1 is set to 10 to transmit the bad Logic Gate map to the terminal device 2 of the driving design company 2 at the terminal. The bad logic gate map is shown on the display portion 21 of the device 20 (see Fig. 2, the bad logic gate map of Fig. 2 is viewed as 2036-8152-PF; Ahddub 1363257 29). In addition, this bad logic gate map can be fabricated on individual wafers or on the entire wafer. Then, the circuit design company 2 can check whether the bad logic is closed, whether it is caused by a certain cause (such as foreign matter mixing, etc.) in the process of each wafer. By transmitting the reason to the wafer fabrication company 3 It can make more efficient improvements to its order and improve the yield (good ratio) in the production of semiconductor integrated circuits. Next, a modification of the navigation window 2 5 (see FIG. 2) will be described with reference to Fig. 15 (please refer to Fig. 1 as appropriate). In Fig. 15, (a) shows the navigation window 25a before the correction resource parameter, and (b) displays the navigation window 25a after the parameter is corrected. In step S1 006 of Fig. 10, the processing unit 15 (see Fig. 1) of the control device ι (see Fig. 1) is like the navigation window 25 (see Fig. 2) to the output unit 13 (see the first Figure) shows the tester that can be used. However, as shown in Figure 15 (a), it is not displayed as such, and the tester list is displayed as in the navigation window 25a. In Figure 15(a), the testers a, B, C and whether they can be used together are displayed. Here, when the operator selects the tester B that is not usable (X: can also be displayed by other symbols), in the next stop, the parameter setting of the tester B is displayed with NG or OK for each resource (cause). Whether it is appropriate to judge the result. In this resource, only the frequency is judged as NG. By correcting the parameter of the frequency, as shown in Fig. 15(b), the judgment result of the frequency becomes OK, and the tester B becomes ready for use. 2036-8152-PF; Ahddub 27 1363257 In addition, when the parameter correction of the frequency is set to 40 nanoseconds and the NG state is set to 40 nanoseconds or more, the value is rewritten to 60 nanometers. Seconds, etc., can be corrected. Thus, by displaying the display such as the navigation window 25a, the operator can easily know the reason why the tester is in an unusable state, and it is easy to make the tester usable by rewriting the parameters and the like. Further, the display of the navigation window 25a, the rewriting of parameters, and the like can be performed on the display unit 21 of the terminal device 20 of the circuit design company 2 that displays a usable tester or the like in step S809 of Fig. 8 . Next, the conversion of the command used on the tester will be described with reference to Fig. 16 (please refer to Fig. 1 as appropriate). Figure 16 is an explanatory diagram showing the conversion of instructions used on the tester. In step S816 of Fig. 8, the test company 1 performs the wafer test using the tester or the like, but at this time, the command used on the tester may not be the terminal device 2 of the circuit design company 2 in Fig. 16 Display section. The command such as "VS" shown above (see Fig. 5) is a command such as "GTL-VS" stored in the memory unit 14 of the control device 1 of the test company i in Fig. 16 in advance. In other words, a command such as "VS" is displayed on the display unit 21 of the terminal device 20 of the circuit design company 2 (see the interactive editor window (3) of Fig. 2, and the processing unit 15 is on the control device 10 of the test company i. When all these instructions are stored in the memory unit 14, the prefix or suffix (here, "GTL-") of w characters or more (here, ^ characters) can be attached. Then, at the end of the tester, The slanting tester Λ ^ attaches the so-called word 2036-8152-PF to all instructions; Ahddub 1363257 first or suffix, interpret and execute these instructions in advance. Thus, even if "vs", "ST〇p" and other instructions are in a certain In the stylized ^, the reserved words (predetermined instructions) representing other meanings, etc., in the tester, H interprets these instructions without any hindrance, and executes the test program. The characters of the right or the suffix can be avoided. The content of the reserved words can be determined by freely combining English symbols, hyphens (_), etc. Next, while referring to the 17th circle, the selection of the function definition used on the tester is explained (please refer to the figure! Etc.) Figure 17 is a flow chart, display and letter For example, as described above, in step S813 of FIG. 8, the test company prepares for the test operation of the tester, and at this time, from the control device ι to the test S- Only the function definition required for the test is stored. As shown in Fig. 17, first, the processing unit 15 of the control device 10 reads the predetermined test program to be used next from the storage unit 14 (step si7〇i). 15 sets i = l (step sl7 〇 3), reads the "description of narration (programma) (step S1705), and determines whether the function definition required for the narration is in the selection function definition file (step S1 707) In addition, the so-called selection function definition slot case refers to the function definition required for the execution of the test program at this time. The collected files are stored in the memory unit 14. If the required function definition is not selected In the function definition file (No in step S1707), the processing unit 15 extracts the function definition from the function definition library of the storage unit 14 and registers it in the selection function definition file (steps s17 and 9). The so-called function definition library' means pre- Many functions are defined and used to define the database that is collected together. 2036-8152-PF; Ahddub 29 1363257 • If the required function is defined in the selection function definition file (Yes in step S1707) and After the step s17a, the processing unit 15 determines whether the narration is the last line (step S1 711). If the narration is not the last line (No in step 31), the processing unit 15 sets i = i + i (Step S1713), the process returns to step s17. 。 5. If the description is the last line (Yes in step 81711), the processing unit 15 ends the process. Thus, the control device 10 collects a call-by-function definition library from a number of function definitions (for example, the number of function definitions is 1 000), and creates a function definition that only collects the execution of the test program at this time (for example, a letter) The selection function defines a library with a number of 2〇〇) and then passes it to the tester. The 'tester' can store only the required function definitions without storing unnecessary function definitions, thus reducing the extra burden and the like, and reducing the processing time required for testing. In addition, the selection (extraction) of the function definition required for 'here' and the use are performed when the test is performed using the tester, but in addition to this, the aforementioned virtual test may be performed (step S811 of Fig. 8). Time to proceed. The above elements have been described as an implementation form, but the form of the present invention is not limited thereto. For example, the present invention can be applied not only to a semiconductor integrated circuit such as an LSI but also to all devices using semiconductor technology such as a memory. In addition, in terms of DFT, not only SCAN technology but also other technologies such as BIST (Built-In Self-Test) can be used. Furthermore, when each device transmits and receives various data through the network 4, it can not only borrow 2036-8152-PF; Ahddub 30 1363257 can use a password and other technologies to improve the security level by encryption and the like. In addition, the specific structures of the hardware, the flowchart, and the like can be appropriately changed without departing from the spirit of the invention. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a view showing the overall configuration of a semiconductor integrated circuit development support system. Fig. 2 is a screen display example of the display unit in the terminal device 2 of the circuit design company 2. Figure 3 is a table showing an example of the specifications of each tester. Figure 4 shows the basic structure common to testers of various modes. Figure 5 shows an example of the tester language TL. Fig. 6(a) is a schematic diagram showing a semiconductor integrated circuit of the RTL method. Fig. 6(b) is a schematic view showing a state in which the multiplexer 62 is attached to each of the flip-flops 61 when the SCAN technique is used in the semiconductor integrated circuit L shown in Fig. 6(a). Figure 7 shows the structure of the reticle shared wafer. Fig. 8 is a flow chart showing the overall operation flow related to the semiconductor integrated circuit development support system 1 000. Figure 9 is a flow chart showing the processing when the tester resources are extracted. Fig. 10 is a flowchart showing the processing when the tester is searched for in the step S8〇8 of Fig. 8 and the cost is calculated. Fig. 11 is a flow chart showing the processing when calculating the increase rate of the crystal area when the DFT using the SCAN technique is calculated. Figure 12 is a flow chart showing the test cost calculation using the reticle sharing 31 - PF; Ahddub 1363257 processing. Fig. 13 is a flowchart showing the arithmetic processing of the test cost of the entire mask shared photomask S in the step sl2〇4 of Fig. 12. Figure 14 is a flow chart showing the processing when a bad logic gate map is created. Fig. 15(a) shows a navigation window 253 before the correction parameter, and the first view shows the navigation window 253 after the correction parameter. Figure 16 is a diagram showing the transformation of the instructions used on the tester. Figure 17 is a flow chart showing the processing associated with the selection of the function definition. [Main component symbol description] 1~ test company; 2~ circuit design company; 3~ wafer fabrication company; 4~ network; 10~ control device; 11~ communication part; 12~ input part; 13~ output part; ~ memory part; 15~ processing part; 16~ memory; 20~ terminal device; 21~ display part; 22~ authentication window; 2 5~ navigation window; 23~ interactive editor window; 25a~ navigation window; 2 4 ~ test pattern window; 26 ~ virtual test window; 27 ~ design window; 28 ~ common reticle window; 29 ~ bad logic gate mapping window. 3 0 ~ terminal device; 50 ~ tester code; 51 ~ component Power supply; 52~C language form; 2036-8152-PF; Ahddub 32 1363257
61 ~正反器; 6 3〜測試電路線; 65〜組合電路; 6 7〜組合電路; 69〜輸出端子; 71 ~放大圖, 30卜電源單元; 303〜比較器; 305〜時序產生器; 307〜DC測試電路; 623~OR邏輯閘; 62~多工器 64〜掃描電路線; 6 6〜時鐘線; 68〜輸入端子; 7 0〜晶圓, 300~測試器; 302〜驅動器; 304~圖案產生器; 306~控制器; 621,622~AND 邏輯閘 L〜半導體積體電路; 1 000〜半導體積體電路開發支援系統。61 ~ forward / reverse; 6 3 ~ test circuit line; 65 ~ combination circuit; 6 7 ~ combination circuit; 69 ~ output terminal; 71 ~ enlarged view, 30 power supply unit; 303 ~ comparator; 305 ~ timing generator; 307~DC test circuit; 623~OR logic gate; 62~ multiplexer 64~ scan circuit line; 6 6~clock line; 68~ input terminal; 7 0~ wafer, 300~ tester; 302~driver; ~ pattern generator; 306~ controller; 621, 622~AND logic gate L~ semiconductor integrated circuit; 1 000~ semiconductor integrated circuit development support system.
2036-8152-PF;Ahddub 332036-8152-PF; Ahddub 33
Claims (1)
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2005/011798 WO2007000806A1 (en) | 2005-06-28 | 2005-06-28 | Semiconductor integrated circuit development support system |
Publications (2)
Publication Number | Publication Date |
---|---|
TW200712810A TW200712810A (en) | 2007-04-01 |
TWI363257B true TWI363257B (en) | 2012-05-01 |
Family
ID=37595070
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW095123119A TW200712810A (en) | 2005-06-28 | 2006-06-27 | Semiconductor integrated circuit development support system |
Country Status (2)
Country | Link |
---|---|
TW (1) | TW200712810A (en) |
WO (2) | WO2007000806A1 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI619955B (en) * | 2012-11-19 | 2018-04-01 | 泰瑞達公司 | Non-transitory computer-readable storage medium,method and test apparatus for debugging in a semiconductor device test environment |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8146061B2 (en) | 2007-12-12 | 2012-03-27 | Via Technologies, Inc. | Systems and methods for graphics hardware design debugging and verification |
CN116774990B (en) * | 2023-08-25 | 2023-11-28 | 合肥晶合集成电路股份有限公司 | Product program management system and management method for semiconductor machine |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3569237B2 (en) * | 2001-03-09 | 2004-09-22 | エー・テイー・イー・サービス株式会社 | Electronic device test means information providing system |
JP2002350499A (en) * | 2001-05-30 | 2002-12-04 | Ando Electric Co Ltd | Test mediating system for semiconductor integrated circuit and test mediating method for it |
JP2003150661A (en) * | 2001-11-15 | 2003-05-23 | Hitachi Ltd | Circuit simulation method |
-
2005
- 2005-06-28 WO PCT/JP2005/011798 patent/WO2007000806A1/en active Application Filing
-
2006
- 2006-06-23 WO PCT/JP2006/312654 patent/WO2007000953A1/en active Application Filing
- 2006-06-27 TW TW095123119A patent/TW200712810A/en not_active IP Right Cessation
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI619955B (en) * | 2012-11-19 | 2018-04-01 | 泰瑞達公司 | Non-transitory computer-readable storage medium,method and test apparatus for debugging in a semiconductor device test environment |
US9959186B2 (en) | 2012-11-19 | 2018-05-01 | Teradyne, Inc. | Debugging in a semiconductor device test environment |
Also Published As
Publication number | Publication date |
---|---|
TW200712810A (en) | 2007-04-01 |
WO2007000953A1 (en) | 2007-01-04 |
WO2007000806A1 (en) | 2007-01-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Marinissen et al. | On IEEE P1500’s standard for embedded core test | |
JP4074738B2 (en) | Event-based semiconductor test system and LSI device design test system | |
JP4058252B2 (en) | IC design verification method | |
JP5188580B2 (en) | Method and apparatus for testing a system on chip with parallel and serial access | |
JP5244196B2 (en) | Method and apparatus for describing components adapted to dynamically modify scan paths for system-on-chip testing | |
JP5184645B2 (en) | Method and apparatus for testing a system on chip with parallel and serial access | |
CN112444731B (en) | Chip testing method and device, processor chip and server | |
JP2008134824A (en) | Power consumption analysis method and program | |
TWI363257B (en) | ||
Vock et al. | Challenges for semiconductor test engineering: a review paper | |
Appello et al. | On the automation of the test flow of complex SoCs | |
TWI363185B (en) | ||
Raik et al. | Constraint-based hierarchical untestability identification for synchronous sequential circuits | |
Sarson et al. | Use models for extending IEEE 1687 to analog test | |
Das et al. | Implementation of a testing environment for digital IP cores | |
Dey et al. | Design for testability techniques at the behavioral and register-transfer levels | |
Huhn et al. | Integrated Circuits | |
Jervan | High-Level Test Generation and Built-In Self-Test Techniques for Digital Systems | |
Melani et al. | An integrated flow from pre-silicon simulation to post-silicon verification | |
Rajsuman | Extending EDA environment from design to test | |
Aderholz et al. | Virtual Test Development Using Pre-Silicon Verification Environment | |
Gregerson et al. | Advances in architectures and tools for FPGAs and their impact on the design of complex systems for particle physics | |
CN118962398A (en) | Chip testing method and device based on FPGA | |
Ong | Test Automation for Software Quality Assurance: A comparative study of hardware and software test automation | |
Portolan et al. | A common language framework for next-generation embedded testing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
MM4A | Annulment or lapse of patent due to non-payment of fees |