TW421792B - Built-in self-test circuit of memory device - Google Patents
Built-in self-test circuit of memory device Download PDFInfo
- Publication number
- TW421792B TW421792B TW88104465A TW88104465A TW421792B TW 421792 B TW421792 B TW 421792B TW 88104465 A TW88104465 A TW 88104465A TW 88104465 A TW88104465 A TW 88104465A TW 421792 B TW421792 B TW 421792B
- Authority
- TW
- Taiwan
- Prior art keywords
- address
- test
- memory
- built
- self
- Prior art date
Links
Landscapes
- Tests Of Electronic Circuits (AREA)
Abstract
Description
經濟部中央標準局員工消費合作社印製 421792 A7 4430twf.doc/008 β 了 五、發明説明(() 本發明是有關於一種記憶體之測試硬體,且特別是有 關於一種記憶體之內置自行測試電路’此內置自行測試電 路與此記憶體係嵌於同一個晶片上。 在記憶體的製造過程中,晶片或覃幕上的灰塵微粒, 刮,以及閘氧化層針孔等所引起的製程相關.錯誤,常易使 記億體受到影響,而導致不當之開路或短路。目前已發展 出許多用來測試記憶體是否正常運作的方法或硬體。然 而,當積體電路技術日新月異之際,記憶體功能也顯得越 來越複雜。也因此,測試記憶體功能所需的時間也相對增 加,而其測試電路也相形複雜。 爲了解決上述問題,已發展出記億體之內置自行測試 (Built-in self-test ; BIST)技術。在進行記憶體內置自行測 試技術的測試時,會有一種演繹被轉換爲由指令、數據和 位址組成之序列,而應用在記憶體中。傳統上,這種轉換 過程通常是利用有限狀態機器來實施。爲了可以進行全速 度測試,測試時BIST電路的操作會像記憶體那樣迅速。這 個方法的缺點之一,就是有限狀態機器會被修改成特定的 一組測試模式,這些側視模式的複雜程度係取決於這個測 試演繹。當測試演繹的複雜性增加,這個方法就無法適用’ 因爲有限狀態機器在測試時的每個時脈週期中,會變得既 大且慢,以至於無法產生一個測試模式。 有些BIST電路,也就是一般熟知的可編程或可配置 BIST電路,可參考美國專利第5,173,906號,“積體電路的 內置自行測試”,係由Dreibelbis等人所揭露,於1992年。 3 (諳先閱讀背面之注意事項再填寫本頁) 訂- r、. 本紙張又度適用中國國家標準(CNS > A4規格(210X297公嫠) 42Π92 Α7 4430twf.doc/008 B7 經濟部中央標準局員工消費合作社印製 五、發明説明(1) 另一種已揭露的電路’係使用以唯讀記憶體內容定義的微 編碼序列來製造測試模式,且於1993年以p0pyack等人的 美國專利第5,224,1〇1號發佈。在美國專利第5,301,156 號,“嵌入式隨機存取記憶體的可配置自行測試”,Talley 等人揭露一種具有一串聯路徑的電路,此串聯路徑通過電 路的位址、指令和數據部份以輸入一個測試或控制模式, 並輸出結果,而每次掃猫需時幾個時脈週期。 本發明提供一種用以測試記憶體之電路與方法,特別 是用以測試記憶體元件。此電路包括一測試模式產生器、 一內置自行測試輸出緩衝器、一延遲緩衝器、以及一比較 器。測試模式產生器更包括三個模組;一有限狀態機器, 用以產生數據和測試模式;一位址產生器,用以進行位址 運算;以及一預測邏輯,用以根據前一個週期的位址資訊, 來預測下一個時脈週期的位址。 本發明採用一週期路徑中斷技術。將正反器配置於模 式產生器的原始臨界路徑,而將原始臨界路徑切成幾個段 落。因此,模式產生器的時脈週期時間取決於這些線段中 的最長時脈週期時間。然而,當這些正反器配置於原始臨 界路徑之後,則產生延遲現象。爲了校正因加入正反器所 導致的時鐘週期之延遲,在本發明中,採用了一個預測邏 輯,配置於兩個正反器之間。由此,而模式產生器的時間 週期時間可以有效地減少,而正反器反引起之延遲問題亦 得以解決。 圖式之簡單說明 - 4 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) (請先閲讀背面之注意事項再填寫本頁) ,裝. -訂 ♦ 經濟部中央標隼局員工消費合作社印裝 42Π92 A7 4430twf.doc/00S 37 五、發明説明(々) 爲讓本發明之上述和其他目的、特徵、和優點能更明 顯易懂,下文特舉較佳實施例,並配合所附圖式,作詳細 說明如下: 第1圖繪示用以測試記憶體元件,一種內置自行測試 電路的方塊圖; 第2(a)圖至第2(c)圖係繪示根據本發明內置自行測試 電路中,模式產生器的演進示意圖;其中 第2(a)圖係爲一單純模式產生器; 第2(b)圖係爲再定時之後的模式產生器;以及 第2(c)圖係爲一種高速模式產生器; 第3圖繪示一種預測邏輯之跨步運算的狀態變化圖; 以及 第4圖繪示一種預測邏輯之暫存器傳送級。 圖式標記說明: 10、2〇、30 :位址暫存器 實施例 在進行一記憶體元件的測試時,測試運算通常是由高階 程式語言寫成。考慮對一包括有A行與C列的記憶體元件 對行測試,其跨步運算包括四個步驟。在第一步驟中,‘0’ 會寫入每個記憶胞中。第二步驟則以正向的方式,對整個 記憶體陣列進行讀出-寫入-讀出的動作,也就是說,從第 一行的第一個記憶胞到最後一行的最後一個記憶胞,在每 個記憶胞中,進行所請的讀出(〇)-寫入(1)-讀出(1)操作。這 個讀出-寫入-讀出的操作,可以顯示出記憶胞的潛在動態 5 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) (請先閱讀背面之注意事項再填寫本頁) 3 許 A7 421792 4430twf.doc/008 B7 五、發明説明(仏) (請先閲讀背面之注意事項再填寫本頁) 缺陷。第三步驟從最後一個記憶胞,經過所有的記憶胞, 而走回第一個記憶胞,並在每個記憶胞中,進行讀出(1)-寫入(〇)-讀出(〇)操作。第四ίέι步驟係檢視在起初的三個步 驟進行後,每個記憶胞的値是否還是零。 此跨步運算可編譯爲一高階語言程式於此舉例如下。 { /*— step 1: write 0 to each cell …*/ foreach_row(i~0; i<R; i++){ foreach_col(j^0; j<C;j++){Write 0 to cell(i, j)} } — step 2: change each cell from 0 to 1 in a forward maimer …*/ foreach_row(i=0; i<R; i++){ foreach_col(j=0; j<C;Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs 421792 A7 4430twf.doc / 008 β V. Description of the invention (() The present invention relates to a test hardware for a memory, and in particular to a built-in Test circuit 'This built-in self-test circuit is embedded on the same chip as this memory system. During the manufacturing process of the memory, the process is caused by dust particles, scratches, and pinholes on the oxide layer of the chip or the screen. . Error, often easy to affect the memory of the billion, resulting in improper open or short circuit. At present, many methods or hardware have been developed to test the normal operation of the memory. However, as the integrated circuit technology changes with each day Memory functions are also becoming more and more complicated. Therefore, the time required to test the memory functions is relatively increased, and its test circuit is relatively complicated. In order to solve the above problems, the built-in self-test of the memory has been developed (Built -in self-test; BIST) technology. When performing the test of the built-in self-test technology of the memory, a deduction is converted into a finger , Data and address sequence, and applied to the memory. Traditionally, this conversion process is usually implemented using a finite state machine. In order to perform full speed testing, the BIST circuit operates like a memory during the test Quick. One of the disadvantages of this method is that the finite state machine will be modified into a specific set of test modes. The complexity of these side-view modes depends on the test interpretation. When the complexity of the test interpretation increases, this method cannot Applicable 'because finite state machines become large and slow in each clock cycle during testing, so that they cannot generate a test mode. Some BIST circuits, which are generally known as programmable or configurable BIST circuits, can Refer to US Patent No. 5,173,906, "Built-in Self-Test of Integrated Circuits", disclosed by Dreibelbis et al., 1992. 3 (谙 Please read the notes on the back before filling this page) Order-r ,. This paper Applicable to Chinese National Standards (CNS > A4 specification (210X297 male)) 42Π92 Α7 4430twf.doc / 008 B7 Central Ministry of Economic Affairs Printed by the Consumer Bureau of the Prospective Bureau. 5. Description of the invention (1) Another type of circuit has been disclosed that uses microcoded sequences defined by the contents of read-only memory to create test patterns. Issued No. 5,224,101. In U.S. Patent No. 5,301,156, "Configurable Self-Test of Embedded Random Access Memory", Talley et al. Disclosed a circuit with a series path that passes through the bits of the circuit. Address, instruction and data part to input a test or control mode and output the result, and each scan takes several clock cycles. The invention provides a circuit and method for testing memory, especially for Test memory components. This circuit includes a test pattern generator, a built-in self-test output buffer, a delay buffer, and a comparator. The test pattern generator further includes three modules; a finite state machine to generate data and test patterns; a bit generator to perform address calculations; and a prediction logic to use the bits from the previous cycle Address information to predict the address of the next clock cycle. The invention uses a periodic path interruption technique. The flip-flop is configured on the original critical path of the mode generator, and the original critical path is cut into several steps. Therefore, the clock cycle time of the pattern generator depends on the longest clock cycle time among these segments. However, when these flip-flops are placed behind the original critical path, a delay occurs. In order to correct the delay of the clock cycle caused by adding the flip-flops, in the present invention, a prediction logic is used, which is arranged between the two flip-flops. Therefore, the time cycle time of the mode generator can be effectively reduced, and the delay caused by the flip-flop is also solved. Brief description of the drawings-4 This paper size is applicable to the Chinese National Standard (CNS) A4 size (210X297 mm) (Please read the precautions on the back before filling this page), and install.-Order ♦ Staff of the Central Bureau of Standards, Ministry of Economic Affairs Printed by a consumer cooperative 42Π92 A7 4430twf.doc / 00S 37 V. Description of the Invention (々) In order to make the above and other objects, features, and advantages of the present invention more comprehensible, the following exemplifies preferred embodiments and cooperates with all The drawings are described in detail as follows: Figure 1 shows a block diagram of a built-in self-test circuit for testing memory components; Figures 2 (a) to 2 (c) show built-in Schematic diagram of the evolution of a pattern generator in a self-test circuit; Figure 2 (a) is a simple pattern generator; Figure 2 (b) is a pattern generator after retiming; and Figure 2 (c) FIG. 3 is a high-speed pattern generator; FIG. 3 is a state change diagram of a step operation of a prediction logic; and FIG. 4 is a register transfer stage of a prediction logic. Description of the graphical symbols: 10, 20, 30: Address register Example When testing a memory component, the test operation is usually written in a high-level programming language. Consider a memory element pair-wise test that includes rows A and C. The step-by-step operation includes four steps. In the first step, '0' is written to each memory cell. The second step is to perform a read-write-read operation on the entire memory array in a positive manner, that is, from the first memory cell in the first row to the last memory cell in the last row. In each memory cell, the requested read (0) -write (1) -read (1) operation is performed. This read-write-read operation can show the potential dynamics of the memory cell. 5 The paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) (Please read the precautions on the back before filling this page. ) 3 Xu A7 421792 4430twf.doc / 008 B7 V. Description of the Invention (仏) (Please read the notes on the back before filling this page) Defects. The third step goes from the last memory cell, passes through all the memory cells, and walks back to the first memory cell, and in each memory cell, reads out (1) -writes (〇)-reads out (〇) operating. The fourth step is to check whether the memory of each memory cell is still zero after the first three steps are performed. This step operation can be compiled into a high-level language program. Here is an example. {/ * — Step 1: write 0 to each cell… * / foreach_row (i ~ 0; i <R; i ++) {foreach_col (j ^ 0; j <C; j ++) {Write 0 to cell (i, j) }} — Step 2: change each cell from 0 to 1 in a forward maimer… * / foreach_row (i = 0; i <R; i ++) {foreach_col (j = 0; j <C;
Read 0 from cell(i3 j)Read 0 from cell (i3 j)
Write 1 to cell(i, j);Write 1 to cell (i, j);
Read 1 from cell(i, j) } 經濟部中央標準局貝工消費合作社印製 step 3: change each cell from 1 to 0 in a backward manner ---*/ foreach—row(i=7?-l; i>=0; i--){ foreach_col(j=(C-l); j>〇; j—Read 1 from cell (i, j)} Printed by Shellfish Consumer Cooperative, Central Bureau of Standards, Ministry of Economic Affairs step 3: change each cell from 1 to 0 in a backward manner --- * / foreach—row (i = 7? -L ; i > = 0; i-) {foreach_col (j = (Cl); j >〇; j—
Read 1 fromcell(ij)Read 1 fromcell (ij)
Write 0 to cell(ij);Write 0 to cell (ij);
Read 0 from cell(i, j) } 6 本紙張尺農適用中國國家標準(CNS ) A4規格(21 297公釐) A7 B7 經濟部中央標準局員工消費合作社印裝 421792 4430twf.doc/008 五、發明説明(s ) /*…step 4: read 0 to each cell …V foreach_row(i=0; i</?; i++){ foreach_col(j=0; j<C; j++){Read 0 to cell(i, j)} 內置自行測試電路 本發明揭露一種內置自行測試電路。如第1圖中所示, 內置自行測試電路包括一模式產生器、一內置自行測試電 路(built-in self-test circuit ; BIST)輸出緩衝器(BOB)、一 延遲緩衝器、以及一比較器。 模式產生器係爲一有限狀態機器,其可產生一與上述 測試運算相關之測試模式序列。bob是一個位在BIST電 路與記憶體元件之間的緩衝器,例如是一個靜態隨機存取 記憶體(SRAM),用以排除由BIST電路引起的位勢特性退 化現象。延遲緩衝器係一先進先出暫存器,位於模式產生 器與比較器之間。主要是爲了計算記憶體讀取指令的預期 反應。延遲時脈週期則取決於記憶體READ操作的延遲時 間。比較器是一個組合邏輯,用以比較記憶體測試後之輸 出反應與此預期反應,以決定記憶體功能操作是否正常’ 其中預期反應係由延遲緩衝器所提供。 在BIST電路的四個組件中,楔式產生器爲主要決定時 脈週期時間的組件。下文將介紹時序臨界路徑的中斷技 術。 第2(a)圖至第2(c)圖繪示根據本發明第1圖中所示的內 7 本紙張尺度適用中國國家標準(CNS ) A4规格(210X297公釐) (請先閲讀背面之注意事項再填寫本頁)Read 0 from cell (i, j)} 6 This paper ruler applies Chinese National Standard (CNS) A4 (21 297 mm) A7 B7 Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs 421792 4430twf.doc / 008 V. Description of the invention (s) /*...step 4: read 0 to each cell… V foreach_row (i = 0; i < / ?; i ++) {foreach_col (j = 0; j <C; j ++) {Read 0 to cell ( i, j)} Built-in self-test circuit The present invention discloses a built-in self-test circuit. As shown in Figure 1, the built-in self-test circuit includes a pattern generator, a built-in self-test circuit (BIST) output buffer (BOB), a delay buffer, and a comparator. . The pattern generator is a finite state machine that can generate a sequence of test patterns related to the above-mentioned test operations. bob is a buffer located between the BIST circuit and the memory element, such as a static random access memory (SRAM), to eliminate the potential characteristics degradation caused by the BIST circuit. The delay buffer is a first-in, first-out register, located between the pattern generator and the comparator. The main purpose is to calculate the expected response to a memory read instruction. The delay clock period depends on the delay time of the memory READ operation. The comparator is a combinational logic that compares the output response after the memory test with the expected response to determine whether the memory function and operation are normal. The expected response is provided by the delay buffer. Among the four components of the BIST circuit, the wedge generator is the component that mainly determines the clock cycle time. Interrupt techniques for timing critical paths are described below. Figures 2 (a) to 2 (c) show the inner paper size shown in Figure 1 of the present invention. The paper size applies to the Chinese National Standard (CNS) A4 specification (210X297 mm) (Please read the back (Please fill in this page again)
421T92 A7 4430twf.doc/〇〇8 gy 五、發明説明(6) 置自fj測試電路中,一種模式產生器的演進圖。第.2(a)圖 繪示一種單純模式產生器的結構。如前所述,模式產生器 係負責產生測試模式,其中每個測試模式包括數據、指令、 和位址等部份。模式產生器因而可分成兩個模組,也就是, 一有限狀態機器,標記以,以及一位址產生器,標 記以⑼。如圖中所示,模組產生測試模式的指 令部份與數據部份,以及一個位址指令,標記以 comm㈣心位址指令係用以驅動模組』, 其係負責產生每個測試模式的位址部份。在本實施例中, 有九個位址指令用來符合各種不同的測試運算。這九個位 址指令包括 N〇_CHANGE、INIT_ADDRESS、LAST_ADDRESS、 NEXT CELL 、 PREVIOUS CELL 、 Y NEXT CELL 、 — · _ — 一 Y_PREVIOUS_CELL ' CROSS—CELL ' 和 RECALL° 根據每個位址 指令,模組«會於每個時脈週期更新位址暫存器10 的値。在這個架構中,可以看到時序臨界路徑起始於位址 暫存器10的輸出,通過模組户«的組合部份, 且最後終結於位址暫存器10的輸入。此外,模式產生器更 包括用於有限狀態機器的原狀態暫存器,標記以 經濟部中央標準局貝工消費合作社印聚 (請先閲讀背面之注意事項再填寫本頁) pg_state。 再定時變換被提出來作一個有效的時序優選裝置。然 而,這在設計上並不適用,因爲臨限路徑實際上會形成一 個環路。第2(b)圖繪示一個再定時變換的示範。再定時位 址暫存器10向後搬移越過模組⑼。在此種再定時形式 中,時序臨界路徑會有些許改變,但時脈週期時間,仍然 8 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公f A7 B7 421T92421T92 A7 4430twf.doc / 〇〇8 gy Fifth, the invention description (6) is placed in the fj test circuit, an evolution diagram of a pattern generator. Fig. 2 (a) shows the structure of a simple pattern generator. As mentioned earlier, the pattern generator is responsible for generating test patterns, where each test pattern includes data, instructions, and addresses. The pattern generator can thus be divided into two modules, that is, a finite state machine, marked with, and a bit generator, marked with ⑼. As shown in the figure, the module generates the command part and data part of the test mode, and an address command, which is marked with the Comm address address command system to drive the module. "It is responsible for generating each test mode. Address part. In this embodiment, there are nine address instructions for conforming to various different test operations. The nine address instructions include NO_CHANGE, INIT_ADDRESS, LAST_ADDRESS, NEXT CELL, PREVIOUS CELL, Y NEXT CELL, — · _ — a Y_PREVIOUS_CELL 'CROSS — CELL' and RECALL ° According to each address instruction, the module « The address register 10 is updated every clock cycle. In this architecture, it can be seen that the timing critical path starts from the output of the address register 10, passes through the combination part of the module user «, and finally ends at the input of the address register 10. In addition, the pattern generator also includes the original state register for the finite state machine, which is marked with the print from the Central Standards Bureau of the Ministry of Economic Affairs of the Bayer Consumer Cooperative (please read the precautions on the back before filling out this page) pg_state. Retiming transformation is proposed as an effective timing optimization device. However, this is not applicable by design because the threshold path actually forms a loop. Figure 2 (b) shows an example of retiming transformation. The retiming address register 10 moves backwards past the module ⑼. In this type of retiming, the timing critical path will be slightly changed, but the clock cycle time is still 8 This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 male f A7 B7 421T92
4430t\vf.doc/00S 五、發明説明(1 ) 主要取決於通過⑼與的組合部份之傳播延遲 的總和。 本發明技術的動機係利用在模組到 之 間,以及㈣到之間置入暫存器,以中斷循環 路徑。這個構想引出了如第2(c)圖所示的一個新架構。除 了兩個組合模組PGjV w與_4_g⑼之外,新架構中還有四 個暫存器與另一個組合式邏輯。這四個暫存器包括原狀態 暫存器pgjiaie ;原胞位址暫存器10 ;暫存器 ,其係表現產生自尸的位址指 令;以及暫存器心,其係表現被單一時脈週 期延遲的一個胞位址。之所以會延遲是因爲在訊號Mdrea 所加入的暫存器。 這個新加入的組合模組,稱之爲預測邏輯,是本技術非 常重要的部份。其目的係根據前一個時脈週期的位址資 料,預測下一個時脈週期的位址。依此,時序改善形式與 原來的形式相當。 . 假定在起初的六個時脈週期,原模式產生器的胞位址爲 Al、A2、A3、A4、A5、和A6。請注意這些位址並非一定 要連續不斷。在高速形式的三個監控點訊號量値、暫存器 、預測邏輯的輸出、以及暫存器中皆列 於表1中。 9 本紙張尺度適用中國图家榡準(CNS > A4規格(210X297公釐) (請先閲讀背面之注意事項再填寫本頁) -訂- ri\ 經濟部中央標準局員工消費合作社印製 42Π92 A7 4430twf.doc/008 B7 五、發明説明(1 ) 時脈週期 1 2 3 4 5 6 原始形式 位址 Α1 Α2 A3 Α4 Α5 Α6 局速 延遲位址 _ Α1 Α2 A3 Α4 Α5 預測位址 — A3 Α4 Α5 Α6 Α7 位址 Α2 A3 Α4 Α5 Α6 表1 (請先閲讀背面之注意事項再填寫本頁) -装· 舉例來說,在第二時脈週期中,當前位址爲A2,而預 測邏輯的輸入則爲延遲的一個A1。以先前所述預測邏輯的 功能描述爲基礎,其輸出應該是下一個時脈週期的位址, 爲A3。這個位址在輸入模組PGjw之前,會被另一個暫 存器所延遲,因此,輸入至/的位址仍是當前位址 A2,其與模式產生器原始形式中者相同。在這個例子中’ 預測邏輯需要由A1來預測A3。這是可以達成的,因爲在 測試運算之下的位址變化序列是經過事先定義的。以下顯 示預測邏輯是如何執行的。較複雜的測試演繹之執行方法 也可以類似的方式推導得之。 爲了能在不喪失一般性的前提下簡化說明’假定記憶體 陣列係爲一個具有N個記憶胞的一維陣列。半跨步運算的 狀態變換圖係顯示於第4圖中’其中標記在電晶體上的變 數a係表示胞位址。除了 START與END狀態之外,圖中 還有五個狀態,也就是S1、S2、S3、S4、和S5。每個狀 10 本紙張尺度適用中國國家標準(CNS ) A4规格(21〇><297公瘦)4430t \ vf.doc / 00S V. Description of the invention (1) It mainly depends on the sum of the propagation delays of the combined parts through AND. The motivation of the technology of the present invention is to use a register between the modules and between them to interrupt the cycle path. This idea leads to a new architecture as shown in Figure 2 (c). In addition to the two combination modules PGjV w and _4_g⑼, there are four registers and another combination logic in the new architecture. These four registers include the original state register pgjiaie; the original cell address register 10; the register, which represents the address instruction generated from the corpse; and the register heart, which represents the time when the performance is single. One cell address of the pulse cycle delay. The delay is due to the register that was added to the signal Mdrea. This newly added combination module, called prediction logic, is a very important part of this technology. The purpose is to predict the address of the next clock cycle based on the address information of the previous clock cycle. Accordingly, the timing improvement form is equivalent to the original form. Assume that in the first six clock cycles, the cell addresses of the original pattern generator are Al, A2, A3, A4, A5, and A6. Please note that these addresses do not have to be continuous. The signal quantities of the three monitoring points in the high-speed form, the register, the output of the prediction logic, and the register are listed in Table 1. 9 This paper size applies to China Tujia Standard (CNS > A4 size (210X297mm) (Please read the precautions on the back before filling out this page)-Order-ri \ Printed by the Central Consumers Bureau of the Ministry of Economic Affairs Consumer Cooperatives 42Π92 A7 4430twf.doc / 008 B7 V. Description of the invention (1) Clock cycle 1 2 3 4 5 6 Original address A1 Α2 A3 Α4 Α5 Α6 Speed-delayed address_ Α1 Α2 A3 Α4 Α5 Predicted address — A3 Α4 Α5 Α6 Α7 Address Α2 A3 Α4 Α5 Α6 Table 1 (Please read the precautions on the back before filling this page)-Equipment · For example, in the second clock cycle, the current address is A2, and the prediction logic is The input is a delayed A1. Based on the functional description of the prediction logic described earlier, its output should be the address of the next clock cycle, which is A3. This address will be input by the other module PGjw before it is input. The register is delayed, so the address input to / is still the current address A2, which is the same as in the original form of the pattern generator. In this example, the prediction logic needs A1 to predict A3. This can be achieved Because of the test operation The following sequence of address changes is defined in advance. The following shows how the prediction logic is performed. The execution method of more complex test deductions can also be derived in a similar way. In order to simplify the description without losing generality 'Assume that the memory array is a one-dimensional array with N memory cells. The state transition diagram of the half-step operation is shown in Figure 4', where the variable a marked on the transistor represents the cell address. Except In addition to the START and END states, there are five states in the picture, namely S1, S2, S3, S4, and S5. Each paper size of 10 papers applies the Chinese National Standard (CNS) A4 specification (21〇 > < 297 male thin)
、1T 經濟部中央標隼局貝工消费合作社印袋 A7 B7 經濟部中央標準局貝工消費合作社印製 d 21 了92 4430twf.doc/008 五、發明説明(句) 態各產生一個唯一測試指令與數據,如每個週期中所指 示。預測邏輯的暫存器轉換級(RTL)碼係顯示於第4圖中。 綜上所述’本發明提供一種用於測試記憶體元件的內 置自行測試電路。最主要的時脈週期時序主導部份是模式 _生器。模式產生器包括一個有限狀態機器尸GJm、一煙 位址產生器一個預測邏輯、以及三個正反器集組, 也就是’一原胞位址暫存器一位址指令暫存器 ⑽心以及用於被延遲單—時脈週期之胞位址 的一個de/吵。在這方法中,時序臨界路徑會被斷 成、Xjert '和預測邏輯三個段落。在這方法中, 時脈週期時間係藉由、义^、和預測邏輯之間的 路徑來決定的。與單純模式產生器的環路相比較,時脈週 期時間得以有效減少。儘管由於增加之暫存器的配置會發 生一個等待時間,這些暫存器的訊號事先還是被預測邏輯 預計算了一個時脈週期。 合成結果 本發明以一個暫存器傳輸水平(RTL)碼執行之。它被以 測試時記憶體元件的功能模型模擬之,以建立其功能修正 的可信度。合成工具,即設計編譯,係使用於將RTL碼轉 化爲一個淨表列,且利用0.35 wmCMOS技術程式庫進行 邏輯優選。靜態時序分析工具,即設定時間,顯示時脈週 期時間係在2.Sns以下。佈局及佈線工具的一個時序驅動 特徵,即Apo 11 〇,係用以產生佈局。基於這些結果,我們 因而可以作以下的結論:本發明所提出的BIST架構有很 本紙張尺度適用中國國家標準(CNS > A4規格(210X297公釐) ,-β (請先閲讀背面之注意事項再填寫本頁)、 1T Printed bag of the Shellfish Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 Printed by the Shellfish Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs d 21 92 4430twf.doc / 008 V. Description of the invention (sentence) Each state generates a unique test instruction With data as indicated in each cycle. The RTL code of the predictive logic is shown in Figure 4. In summary, the present invention provides a built-in self-test circuit for testing a memory element. The most important part of the clock cycle timing is the mode generator. The pattern generator includes a finite-state machine corpse GJm, a smoke address generator, a predictive logic, and a set of three flip-flops, that is, a 'primary cell address register and a single address instruction register. And a de / noise for the delayed single-clock cycle cell address. In this method, the timing critical path will be broken into three paragraphs, Xjert 'and prediction logic. In this method, the clock cycle time is determined by the path between, meaning, and prediction logic. Compared with the loop of the simple mode generator, the clock cycle time is effectively reduced. Although there is a waiting time due to the increased configuration of the registers, the signals of these registers are pre-calculated by the prediction logic for a clock period. Synthetic results The present invention is implemented as a register transfer level (RTL) code. It is simulated with a functional model of the memory element during the test to establish the credibility of its functional modification. The synthesis tool, design compilation, is used to convert the RTL code into a net list, and uses the 0.35 wmCMOS technology library for logic optimization. The static timing analysis tool, that is, set time, shows that the clock cycle time is below 2.Sns. A timing-driven feature of the place and route tool, Apo 110, is used to generate the place. Based on these results, we can therefore make the following conclusions: The BIST architecture proposed by the present invention has a very large paper size that is applicable to Chinese national standards (CNS > A4 specification (210X297 mm), -β (Please read the precautions on the back first) (Fill in this page again)
421792 A7 4430twf.doc/008 五、發明説明(γ) 大的潛力可以4ΟΟΜΗζ之局速運轉’迨與不具适項技術的 單純實施比較起來,加快了約30%。 在考慮本說明書並實行其中所揭露的發明之後’熟習 此技藝者將發現本發明其他可能的實施例。於此,本說明 書及其實施例係僅用以舉例說明而已’任何熟習此技藝 者,在不脫離本發明之精神和範圍內’當可作各種之更動 舆潤飾'。因此,本發明之保護範圍當視後附之申請專利範 圍所界定者爲準。 (請先閱讀背面之注意事項再填寫本頁〕 訂 經濟部中央橾準局貝工消費合作社印製 本紙張尺度it肖巾( CNS ) Μ酿(21〇><1^楚)421792 A7 4430twf.doc / 008 V. Description of the invention (γ) The large potential can be run at a speed of 400MΗζ, which is about 30% faster than the simple implementation without proper technology. After considering this specification and practicing the invention disclosed therein ', those skilled in the art will find other possible embodiments of the invention. Herein, this description and its embodiments are only for illustration purposes. 'Any person skilled in the art can make various modifications without departing from the spirit and scope of the present invention'. Therefore, the protection scope of the present invention shall be determined by the scope of the appended patent application scope. (Please read the precautions on the back before filling in this page.) Order Printed by the Central Bureau of Standards, Ministry of Economic Affairs, Shellfish Consumer Cooperative, This paper size it shawl (CNS) M brewing (21〇 > < 1 ^ chu)
Claims (1)
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US26221499 | 1999-03-03 |
Publications (1)
Publication Number | Publication Date |
---|---|
TW421792B true TW421792B (en) | 2001-02-11 |
Family
ID=22996647
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW88104465A TW421792B (en) | 1999-03-03 | 1999-03-22 | Built-in self-test circuit of memory device |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW421792B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI646845B (en) * | 2016-05-19 | 2019-01-01 | 晨星半導體股份有限公司 | Conditional access chip, built-in self-test circuit and test method thereof |
-
1999
- 1999-03-22 TW TW88104465A patent/TW421792B/en not_active IP Right Cessation
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI646845B (en) * | 2016-05-19 | 2019-01-01 | 晨星半導體股份有限公司 | Conditional access chip, built-in self-test circuit and test method thereof |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US6318911B1 (en) | Gated clock design supporting method, gated clock design supporting apparatus, and computer readable memory storing gated clock design supporting program | |
TWI221614B (en) | Tester architecture construction data generating method, tester architecture constructing method and test circuit | |
TW563029B (en) | Flexible interrupt controller that includes an interrupt force register | |
US5980092A (en) | Method and apparatus for optimizing a gated clock structure using a standard optimization tool | |
US7506292B2 (en) | Method for clock synchronization validation in integrated circuit design | |
US12039240B2 (en) | Integrated circuit simulation and design method and system thereof | |
JP4200465B2 (en) | Semiconductor integrated circuit design method and design system | |
US10303833B1 (en) | Parallelizing timing-based operations for circuit designs | |
TW421792B (en) | Built-in self-test circuit of memory device | |
JP2000187676A (en) | Logical synthesizing device and computer readable recording medium recording its program | |
US20030154434A1 (en) | Self testing-and-repairing data buffer and method for operating the same | |
JP3667146B2 (en) | Built-in self-test circuit for memory | |
US6175518B1 (en) | Remote register hierarchy accessible using a serial data line | |
TWI288242B (en) | Digital logic test method to systematically approach functional coverage completely and related apparatus and system | |
JP3207109B2 (en) | Scannable last-in first-out register stack | |
JP3725230B2 (en) | LSI function design support device | |
EP1031995A1 (en) | Built-in self-test circuit for memory | |
JP2883465B2 (en) | Electronic computer | |
JP3563223B2 (en) | Register circuit | |
JP3628545B2 (en) | Built-in self-test circuit for memory devices | |
JP3124598B2 (en) | Address generation circuit | |
JPH10340596A (en) | Data storage device and semiconductor memory | |
JP3063694B2 (en) | State machine control circuit | |
JPH11259554A (en) | Timing analysis method for logic circuit, logic synthesis system using the method and record medium for programming and recording the method | |
JP2009145957A (en) | State machine and semiconductor integrated circuit using it |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
GD4A | Issue of patent certificate for granted invention patent | ||
MK4A | Expiration of patent term of an invention patent |