TW384475B - Improved techniques for reducing redundant element fuses in a dynamic random access memory array - Google Patents

Improved techniques for reducing redundant element fuses in a dynamic random access memory array Download PDF

Info

Publication number
TW384475B
TW384475B TW087109644A TW87109644A TW384475B TW 384475 B TW384475 B TW 384475B TW 087109644 A TW087109644 A TW 087109644A TW 87109644 A TW87109644 A TW 87109644A TW 384475 B TW384475 B TW 384475B
Authority
TW
Taiwan
Prior art keywords
spare
components
defective
group
component group
Prior art date
Application number
TW087109644A
Other languages
Chinese (zh)
Inventor
Joerg Vollrath
Original Assignee
Siemens Ag
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US08/884,854 external-priority patent/US5831917A/en
Application filed by Siemens Ag filed Critical Siemens Ag
Application granted granted Critical
Publication of TW384475B publication Critical patent/TW384475B/en

Links

Classifications

    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C29/00Checking stores for correct operation ; Subsequent repair; Testing stores during standby or offline operation
    • G11C29/70Masking faults in memories by using spares or by reconfiguring
    • G11C29/78Masking faults in memories by using spares or by reconfiguring using programmable devices
    • G11C29/785Masking faults in memories by using spares or by reconfiguring using programmable devices with redundancy programming schemes
    • G11C29/787Masking faults in memories by using spares or by reconfiguring using programmable devices with redundancy programming schemes using a fuse hierarchy
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C29/00Checking stores for correct operation ; Subsequent repair; Testing stores during standby or offline operation
    • G11C29/70Masking faults in memories by using spares or by reconfiguring
    • G11C29/78Masking faults in memories by using spares or by reconfiguring using programmable devices
    • G11C29/783Masking faults in memories by using spares or by reconfiguring using programmable devices with refresh of replacement cells, e.g. in DRAMs
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C29/00Checking stores for correct operation ; Subsequent repair; Testing stores during standby or offline operation
    • G11C29/70Masking faults in memories by using spares or by reconfiguring
    • G11C29/78Masking faults in memories by using spares or by reconfiguring using programmable devices
    • G11C29/80Masking faults in memories by using spares or by reconfiguring using programmable devices with improved layout
    • G11C29/812Masking faults in memories by using spares or by reconfiguring using programmable devices with improved layout using a reduced amount of fuses

Landscapes

  • For Increasing The Reliability Of Semiconductor Memories (AREA)

Abstract

A memory array having a first plurality of fuse-sharing redundant elements for replacing defective elements of the memory array. The memory array includes a first fuse, and first group of redundant elements of the first plurality of fuse-sharing redundant elements. The first group of redundant elements share the first fuse as their highest order address fuse. The memory array further includes a second group of redundant elements of the first plurality of fuse-sharing redundant elements. The second group of redundant elements is mutually exclusive with respect to the first group of redundant elements.

Description

經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(,) 發明背景 發明._領域 本發明俗有關半導體裝置的設計和製造。特別是,本 發明傜有關增加記億體電路内之電路密度的一種改良技 術。 相關搜1術説、社_ 於一値記億體電路例如動態隨機存取億體(D R A Μ)或 是概(field)可程式控制的邏輯裝置内,為了定址之目 的滴常將記億體單元(c e 1 1 )配置成行和列。拿實例來説 ,一個典型的動態隨機存取晶Η可能具有六仟四百萬甚 或更多單元,這些單元可以配置成行和列而由字元線和 位元線所定址。因此為求簡潔,在此將不會對習知裝置 中所熟知之動態隨機存取記億體的電路和設計作詳細的 射論 〇 典型的DRAM晶片内,可能會在主陣列的數百萬橱單元 中發現一個或更多個單元是有瑕疵的。設計者通常會提 供備用單元而不是將整個晶Η拋棄,而這些備用單元可 以用來替換那些有瑕疵的單元。於使用期間,以備用單 元取代替換掉的瑕疵單元,因而允許吾人在使用記億體 電路時可以當作其中好像沒有任何瑕疵一般。 若於製造期間發現主記億體陣列中有一個單元是有瑕 疵的,通常會將含有瑕疵單元的整個行或列由備用的行 或列所取代。為易於參閲,以下是將單元的整値行或列 稱為一値元件。為易於了解,以下將參照各列及其替換 本紙張尺度適用中國國家標準(CNS ) Α4規格(210X297公釐) 訂 ν'^ (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(> ) 列而進行討論,然應謹記以下所討論的試題可以等效地 應用於各行及其替換行。 當以備用元件取代陣列中有瑕疵的元件時,習知製程 中的替換技術會於製造期間涉及備用電路之熔絲的設定 和啓動以標示使用的是備用元件而不是主陣列元件之一 。由備用元件所取代之瑕疵主耳聞列元件的定址,是於 替換時間之内藉著設定該備用元件的位址熔絲而指定的 „於某些D R A M s内,啓動熔絲及位址熔絲的值可以分別載 λ到啓動插梢(latch)及位址插梢之内。若啓動插梢所含 的倌標明了應該使用備用元件,則會以備用元件取代位 址插梢所指定的位址上的有瑕疵之主陣列元件。 為了作更詳盡的説明,第1圔展示出一個具有主陣列 102而經毫簡化的DRAM晶Η 100。雖然實際的主陣列通常 具有的元件數目會多得多,圖中所顯示的主陣列1 〇 2只 具有四個列或元件(0-3)以便易於討論。圖中也顯示了 一個可以用來取代任何一個元件0-3的備用列或元件104 。於真實的D R A Μ晶Η内,很明顯地通常會有更多的列或 元件。I比外,主陣列可能包括許多子陣列。不過,為簡 化討論的内容這裡所顯示的DRAM晶Η具有一個含四値列 或元件的主陣列。 為了取代任何一個元件0-3,備用元件104是與兩個位 址位元A 1和A 0有關聯。位址位元的值會為解碼邏輯電路 指出備用元件是取代了主陣列102中的那一個主陣列元件 。第2圖展示習知製程中所用的備用電路圖以指出備用 -4 - 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) --------Γ d裝------訂-----T./m (請先閲讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(4 ) 元件104的替換位址。如第2圖所示,備用電路210包含 一個與解碼邏輯2 0 2耦合備用元件1 0 4。解碼邏輯2 0 2所 檐負的貴任,是以製造期間所設定的啓動及位址熔絲為 基礎而確定是否應該以備用元件104來取代瑕疵之主陣列 元件,S若其答案為是,則得確定是那一個主陣列元件。 第2圖中,啓動晶Η是顯示成啓動晶片EF,而如第1 圖所示的位址晶Η刖顯示成位址晶M F 1和F 0。為了進行 討論而假定於D R A Μ單元1 0 0的品管期間發現主陣列元件 种2是有瑕疵的。此例中,將啓動晶HEF設定成會標示 出應該把備用陣列元件1 0 4拿來替換。在位址晶H F 0不 是設定為形成位元圖案「10」下會設定址晶HF1而指 出以備用元件1 0 4取代主陣列元件2。 於電源開啓期間亦即於D R A Μ晶片的蓮轉時間内,將啓 動熔絲E F的倌載入到第2圔的啓動插梢E L内。此例中, 會將數倌「1」載入到啓動插梢E L内。位址熔絲F 1和F 0 的倌也會分別載入到位址插梢A L 1和A L 0内,致使例如位 址捅梢AL1會儲存數值「1」而位址插梢AL0會儲存數值 「0」。使用插梢是由於它們在運轉期間的讀取速率比 熔絲還快,因而能將DRAM單元的操作速率最佳化。藉著 檢測啓動插梢EL,解碼邏輯202能確定應該以備用元件 1 0 4取代有瑕疵的陣列元件。藉箸檢測位址插梢A L 1和A L 0 ,解碼邏輯202能確定是以備用元件104取代有瑕疵的陣 列元件if 2。 上述習知技術雖在結合第2圖的情況下能適當地取代 -5 - 本紙張尺度適用中國國家標準(CNS ) A4規格(210 X 297公釐) (讀先閱讀背面之注意事項再填寫本頁) .T裝.Printed by the Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Description of the Invention (,) Background of the Invention Invention _ Field This invention relates to the design and manufacture of semiconductor devices. In particular, the present invention relates to an improved technique for increasing the circuit density in a memory circuit. Relevant search 1 theory, agency _ Yu Yiji remember billion body circuits such as dynamic random access billion body (DRA M) or field (programmable logic device), for the purpose of addressing will often remember billion body The cells (ce 1 1) are arranged in rows and columns. For example, a typical dynamic random access chip may have 64 million or even more cells. These cells can be arranged in rows and columns and addressed by word lines and bit lines. Therefore, for the sake of brevity, the circuit and design of dynamic random access memory, which is well-known in conventional devices, will not be discussed in detail here. In a typical DRAM chip, there may be millions in the main array. One or more units were found to be defective in the cabinet unit. Designers often provide spare units instead of abandoning the entire wafer, and these spare units can be used to replace defective units. During use, the defective unit replaced by a spare unit allows us to use it as if there is no defect in it. If a cell in the master billion array is found to be defective during manufacturing, the entire row or column containing the defective cell is usually replaced by a spare row or column. For ease of reference, the entire row or column of cells is referred to below as a stack of components. For easy understanding, the following will refer to each column and its replacement. This paper size applies the Chinese National Standard (CNS) Α4 specification (210X297 mm). Order ν '^ (Please read the precautions on the back before filling this page) Central Standards of the Ministry of Economic Affairs A7 B7 printed by the Bureau ’s Consumer Cooperatives V. Invention Description (>) column for discussion, but it should be kept in mind that the test questions discussed below can be equivalently applied to each line and its replacement line. When a defective component in the array is replaced with a spare component, the replacement technology in the conventional manufacturing process involves the setting and activation of the fuse of the spare circuit during manufacturing to indicate that the spare component is used instead of one of the main array components. The addressing of the defective main ear device replaced by a spare component is specified within the replacement time by setting the address fuse of the spare component. „In some DRAMs, start the fuse and address fuse The value of λ can be loaded into the start pin and the address pin respectively. If the 含 included in the start pin indicates that a spare component should be used, the spare component will replace the bit specified by the address plug. The defective main array element at the address. For more detailed explanation, the first one shows a simplified DRAM chip 100 with the main array 102. Although the actual main array usually has a large number of components Many, the main array 1 shown in the figure has only four columns or elements (0-3) for easy discussion. The figure also shows a spare column or element 104 that can be used to replace any element 0-3 In a real DRA M crystal, it is clear that there will usually be more columns or elements. In addition, the main array may include many sub-arrays. However, to simplify the discussion, the DRAM crystal shown here has One with four queues or The main array of components. To replace any one of components 0-3, spare component 104 is associated with two address bits A 1 and A 0. The value of the address bit will indicate to the decoding logic circuit that the spare component has replaced The main array element in the main array 102. Figure 2 shows the spare circuit diagram used in the conventional process to indicate the spare -4-This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) ---- ---- Γ d ------Order-----T./m (Please read the precautions on the back before filling out this page) Printed by the Consumer Standards Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 5. Description of the Invention (4) The replacement address of the element 104. As shown in Fig. 2, the backup circuit 210 includes a backup element 1 0 4 coupled with the decoding logic 2 0. The responsibility of the decoding logic 2 0 2 is Based on the start-up and address fuses set during manufacturing, determine whether the defective main array element should be replaced by the spare element 104. If the answer is yes, then it must be determined which main array element is. Figure 2 The boot chip is shown as the boot chip EF, and the address chip shown in Figure 1 The address crystals MF 1 and F 0 are displayed. For discussion, it is assumed that the main array element type 2 is found to be defective during the quality control of the DRA M unit 100. In this example, the startup crystal HEF is set to indicate It should be replaced with the spare array element 104. When the address crystal HF 0 is not set to form the bit pattern "10", the address crystal HF1 is set and it is pointed out that the main array element 2 is replaced with the spare element 104. During the power-on period, that is, during the turn-on time of the DRAM chip, the start-up fuse EF is loaded into the start-up pin EL of the second volume. In this example, the number "1" will be loaded into the starting plug EL. The addresses of address fuses F 1 and F 0 are also loaded into address pins AL 1 and AL 0, respectively, so that, for example, address pin AL1 stores the value "1" and address pin AL0 stores the value " 0 ". Pins are used because they can read faster than fuses during operation, which optimizes the operating rate of DRAM cells. By detecting the activation of the pin EL, the decoding logic 202 can determine that the defective array element should be replaced with a spare element 104. By detecting the address pins A L 1 and A L 0, the decoding logic 202 can determine that the defective array element if 2 is replaced with the spare element 104. Although the above-mentioned conventional technology can appropriately replace -5 in combination with Figure 2-This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) (Read the precautions on the back before filling in this paper Page) .T pack.

、1T » 1 ' A7 B7 經濟部中央標準局員工消費合作社印製 五、發明説明 ( 4- ) - 1 1 有 抿 疵 的 主 陣 列 元 件 9 然 而 還 是 有 一 些 缺 點 Ο 藉 箸 實 例 1 1 I ♦ 隨 著 DR A Μ晶 片 容 量 的 增 加 需 要 將 更 大 數 百 的 主 陣 列 1 1 記 億 體 元 件 及 / 或 備 用 元 件 放 進 給 定 尺 寸 的 晶 片 内 〇 不 讀 1 1 倘 * 每 値 熔 絲 都 在 D R A Μ 晶 Η 上 占 用 __. 値 不 明 顯 的 面 聞 1 背 1 稽 使 得 習 知 趨 近 法 中 所 要 求 的 大 數 巨 熔 絲 不 利 地 要 求 之 1 I 更 大 的 鑄 模 (D i e )尺寸以容納所有的熔絲。 '-‘ 意 攀 1 為 了 展 示 之 用 i 若 陣 列 中 有 2 1値元件則每- '個備用 項 • 1 i i 元 件 都 必 需 有 η 橱 位 址 熔 絲 〇 每 一 値 備 用 元 件 也 必 需 有 寫 本 裝 一 値 啓 動 熔 絲 9 以 便 如 上 述 討 論 於 運 轉 期 間 標 示 出 是 否 頁 1 1 應 該 使 用 該 備 用 元 件 Ο 因 此 i 在 習 知 設 計 中 每 一 値 備 用 1 I 7Π 件 必 需 具 有 的 熔 絲 總 數 是 η + 1 〇 若每- -個陣列中是提 1 1 供 有 m 個 備 用 元 件 (具有取代主陣列中的m 個瑕疵元件) 1 訂 則 所 需 要 的 熔 絲 總 數 是 m * ( η + 1) 〇 值 得 熟 悉 習 用 技 術 1 I 的 人 注 的 是 > 若 能 減 少 熔 絲 總 數 則 能 夠 減 少 鑄 模 尺 寸 1 1 、 減 小 DR AM晶 Η 的 尺 寸 ) 及 / 或 增 加 D R A Μ 的 密 度 〇 1 1 於 刖 述 觀 點 中 ) 預 期 有 改 良 技 術 能 為 DR AM陣 列 提 供 備 用 元 件 而 有 利 地 減 小 了 其 施 行 時 所 需 要 的 面 積 * 因 而 能 1 在 給 定 尺 寸 的 晶 片 内 放 置 更 大 數 目 的 主 陣 列 元 件 及 / 或 1 1 備 用 元 件 〇 1 | 發 明_ Μ 述- 1 1 於 一 橱 實 施 例 中 9 本 發 明 偽 有 關 一 種 於 記 億 體 元 件 内 1 | 製 成 許 多 第 一 備 用 元 件 以 簡 少 所 用 位 址 熔 絲 數 百 的 方 法 1 η 11 種 方 法 包 含 確 定 許 多 第 一 備 用 元 件 中 備 用 元 件 的 數 1 1 巨 是 至 少 有 個 〇 若 許 多 6 第 備 用 元 件 中 備 用 元 件 的 數 1 1 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) kl B7 經濟部中央標準局員工消費合作社印裝 五、發明説明( r ) \ 1 是 至 少 有 三 痼 則 這 種 方 法 包 含 藉 由 許 多 第 一 備 用 元 1 1 I 件 形 成 第 一 備 用 元 件 群 而 將 位 址 熔 絲 最 佳 化 的 步 驟 〇 此 1 1 最 佳 化 的 步 驟 另 外 包 含 由 許 多 第 一 備 用 元 件 形 成 第 二 備 -V 請 先 1 1 用 元 件 群 的 步 驟 〇 第 二 備 用 元 件 群 與 第 備 用 元 件 群 之 閲 ik 1 背 1 間 是 η 斥 的 〇 此 最 佳 化 的 步 驟 也 包 含 提 供 一 痼 第 一 熔 絲 1¾ 之 1 熔 絲 注 | 以 供 第 —- 備 用 元 件 群 中 的 各 備 用 元 件 共 用 〇 第 意 事 1 形 成 第 一 備 用 元 件 群 中 各 備 用 元 件 的 最 局 階 位 址 熔 絲 〇 項 1 / 、 於 另 一 m 實 施 例 中 本 發 明 傺 有 關 一 種 記 憶 體 陣 列 , 其 填 寫 本 % 中 有 許 多 共 用 熔 絲 之 第 __* 備 用 元 件 以 取 代 記 億 體 陣 列 中 頁 S. 1 1 的 瑕 疵 元 件 方 此 記 億 體 Π K 陣 列 包 含 一 個 第 一 熔 絲 以 及 由 1 I 許 多 共 用 熔 絲 之 第 一 備 用 元 件 構 成 的 第 一 備 用 元 件 群 〇 1 1 各 第 一 備 用 元 件 群 是 以 共 用 第 一 熔 絲 作 為 它 們 的 最 高 階 1 訂 位 址 熔 絲 Οι 此 記 億 體 陣 列 另 外 包 含 由 許 多 共 用 熔 絲 之 第 1 I 一 備 用 元 件 構 成 元 件 構 成 的 第 二 備 用 元 件 群 〇 第 二 備 用 I 1 元 件 群 與 第 一 備 用 元 件 群 之 間 是 互 斥 的 〇 1 I 於 又 一 個 實 施 例 中 本 發 明 傺 有 關 一 種 電 腦 -施行方法, .丄-、 / 匕' 偽 用 於 將 記 億 體 陣 列 中 許 多 第 一 瑕 疵 元 件 取 代 成 許 多 共 來 1 用 熔 絲 之 第 一 備 用 元 件 中 的 備 用 元 件 〇 m 種 電 腦 -施行 1 | 方 法 包 含 確 定 許 多 第 一 瑕 疵 元 件 中 瑕 疵 元 件 的 數 目 是 否 1 I 小 於 許 多 共 用 熔 絲 之 第 一 備 用 元 件 中 備 用 元 件 的 數 巨 步 1 1 驟 〇 若 許 多 第 一 瑕 疵 元 件 中 瑕 疵 元 件 的 數 百 是 小 於 許 多 1 I 共 用 熔 絡 之 第 一 備 用 元 件 中 備 用 元 件 的 數 巨 , 則 這 種 電 1 1 腦 -施行方法另外包含於許多第- -瑕疵元件中加入許多 1 1 偽 诰 的 瑕 疵 元 件 以 便 將 許 7 多 第 瑕 疵 元 件 中 瑕 疵 元 件 的 1 1 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) A7 Β7 經濟部中央標準局員工消費合作社印製 五、發明説明( b ) - - 1 1 數 百 增 加 到 許 多 共 用 熔 絲 之 第 一 備 用 元 件 中 備 用 元 件 的 1 1 I 數 目 〇 1 1 這 種 電 腦 -施行方法也包含確定許多第- -瑕疵元件中 r-S 請 先 1 1 瑕 疵 元 件 的 數 圈 是 否 至 少 有 三 値 〇 若 許 多 第 一 瑕 疵 元 件 閲 讀 1 背 1 中 瑕 疵 元 件 的 數 百 是 至 少 有 三 痼 則 這 種 電 腦 -施行方法 1¾ 之 1 注 | 包 含 利 用 執 行 由 許 多 第 - 瑕 疵 元 件 形 成 第 —- 瑕 疵 元 件 群 意 事 1 的 步 驟 » 而 將 許 多 第 一 瑕 疵 元 件 中 的 個 別 元 件 取 代 成 許 項 再 ;j 多 第 一 備 用 元 件 中 的 個 別 元 件 〇 第 一 瑕 疵 元 件 群 中 的 各 填 寫 本 1 裝 瑕 疵 元 件 於 最 高 階 位 址 位 元 的 位 置 上 共 用 相 同 的 位 元 值 頁 、«_〆 I I η m 種 取 代 步 驟 另 外 包 含 由 許 多 第 一 瑕 疵 元 件 形 成 第 二 I I 瑕 疵 元 件 群 的 步 驟 〇 第 二 瑕 疵 元 件 群 與 第 一 瑕 疵 元 件 群 I I 之 間 是 互 ff 的 〇 這 種 取 代 步 驟 也 包 含 將 第 一 瑕 疵 元 件 群 I 訂 中 的 個 別 元 件 分 配 到 第 一 備 用 元 件 群 中 的 値 別 元 件 以 便 I | 淮 行 耽 代 的 步 驟 〇 I I 太 發 明 的 m 及 其 他 持 性 將 於 以 下 詳 細 中 參 照 所 附 圔 I I 示 作 更 詳 畫 的 説 明 ,'ΐ 圖 式 簡 述 I 本 發 明 是 m 實 例 而 展 示 的 然 並 不 局 限 於 此 > 所 附 圖 表 I I 的 各 圖 中 相 同 的 參 考 標 碼 所 標 示 的 是 類 似 的 元 件 〇 I I 圖 式 之 簡 単 説 明 I I 第 1 圖 展 示 的 是 經 局 度 簡 化 以 利 討 m 而 具 有 一 m 主 I I 陣列的D R A Μ單 元 〇 1 1 1 第 2 圖 展 示 的 是 習 知 設 計 中 所 用 的 備 用 電 路 圖 以 指 1 1 出 備 用 元 件 的 替 換 位 址 〇 1 I 8 1 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210 X 297公釐) A7 B7 經濟部中央標準局員工消費合作社印製 五、發明説明( 7 ) 1 1 1 第 3 圖 偽 為 了 方 便 討 論 共 用 熔 絲 之 可 能 性 而 展 示 1 1 1 了 五 値 解 釋 用 瑕 疵 元 件 以 及 它 們 的 位 址 〇 1 I 第 4 A HI 展 示 的 是 於 習 知 設 計 中 提 供 - 個 各 具 有 四 個 請 先 1 1 閲 I 位 址 熔 絲 之 備 用 元 件 時 的 所 有 熔 絲 〇 讀 背 1 I 第 4B圖 展 示 的 是 於 替 換 時 間 内 可以如何設定第4Α圖 面 之 1 1 中 的 榕 絲 以 利 替 換 第 3 圖 實 例 中 的 瑕 疵 JQ 件 〇 意 事 1 第 5 圖 展 示 的 是 根 據 本 發 明 —- 個 實 施 例 中 値 共 用 項 再 1 .·*ν 1 \ . 裝 熔 絲 之 備 用 元 件 以 及 供 相 鄰 備 用 元 件 共 用 以 節 省 熔 絲 數 本 頁 1 冃 的 熔 絲 配 置 〇 V✓ 1 第 6 圖 俗 第 5 圔 中 備 用 元 件 的 簡 略 電 路 圔 〇 1 | 第 7 圖 展 示 的 是 可 以 如 何 設 定 第 5 圖 中 備 用 元 件 的 1 | 熔 絲 以 利 替 換 第 3 圖 之 .·» Λ. 刖 述 實 例 中 的 瑕 疵 陣 列 元 件 〇 1 訂 第 8 圖 展 示 的 是 根 據 本 發 明 中 一 個 概 念 於 製 造 期 間 1 I 在 共 用 熔 絲 巨 的 下 用 來 組 合 備 用 元 件 的 方 法 〇 1 1 第 9 圖 展 示 的 是 應 用 第 8 圖 的 方 法 以 便將第4A圖之 1 1 習 知 設 計 中 所 提 供 的 熔 絲 數 巨 最 佳 化 〇 丄 W 第10圖 % 展 示 的 是 根 據 本 發 明 中 一 値 概 念 於 替 換 期 間 1 在 後 續 將 每 一 値 瑕 疵 元 件 分 配 到 共 用 熔 絲 之 備 用 元 件 中 1 的 適 當 備 用 元 件 上 之 百 的 下 用 來 組 合 瑕 疵 元 件 的 方 法 〇 1 I 第11圖 % 展 示 的是將第1 0圖的方法應用 在 第 3 圖 實 例 1 I 中 的 情 況 〇 1 1 第 1 2圖 Λ 展 示 的 是 將 第 1 0圖 的 方 法 m 用 在 一 群 三 値 其 1 I 位 址 分 別是000 1 0101 > 及 1111的瑕 疵 元 件 上 之 情 況 〇 1 1 第 1 3圖 > 展 示 的 是 將 第 9 10圖的方法應用 在 群 」—▲ 値 其 1 1 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) A7 B7 經濟、部中央標準局員工消費合作社印装 五、發明説明( S ) 1 1 位 址 分 別 是 0000 » 000 1 9 00 10 9 0011 9 01 00 9 及 110 1 的 1 1 I 瑕 疵 元 件 上 之 情 況 〇 1 1 發 明 的 詳 細 説 明 請 A, 1 1 現 在 將 參 照 幾 値 展 示 用 的 實 施 例 及 所 附 圖 表 來 說 明 本 聞 讀 1 背 1 發 明 〇 於 下 列 説 明 中 » 設 立 了 很 多 待 定 的 細 節 以 便 提 供 之 1 I 對 本 發 明 的 τη 整 理 解 〇 不 過 對 熟 悉 習 用 技 術 的 人 而 -Λ-. > 意 1 很 明 顯 地 本 發 明 可 以 在 所 附 申 請 專 利 範 圍 的 某 或 全 部 項 再 Li -.Λ' 項 冃 下 奮 行 λ 於 其 他 狀 況 中 t 為 了 不 致 對 本 發 明 的 理 Λ7Ϊ m 填 寫 本 i 诰 成 不 必 要 障 礙 > 故 在 此 不 ~r^~ 對 吾 人 所 熟 知 的 處 理 步 驟 頁 1 1 / 或 結 構 作 詳 細 的 說 明 〇 1 I 一 般 而 言 > 所 發 明 的 熔 絲 節 省 技 術 可 以 同 時 用 於 像 包 1 1 含 DR AM s和同步DRAMs (SDRAMs) 等 動 態 隨 機 存 取 記 億 體 1 訂 (R AM S ) 之 類 記 億 體 積 體 電 路 (ICs )的製造期間、 以及瑕 1 I 疵 主 陣 列 元 件 的 替 換 期 間 〇 其 他 的 1C s包含像櫥可程式 1 1 控 制 的 閘 極 陣 列 (FPG As )之類的邏輯晶Η或是其他邏輯 1 1 晶 Η 〇 於 DR AM的 製 造 期 間 可 以 使 用 熔 絲 節 省 技 術 來 決 0 定 如 何 於 所 提 供 的 備 用 元 件 之 間 共 用 熔 絲 〇 這 是 因 為 通 1 常 在 不 同 型 式 的 DR AM 電 路 中 有 不 同 數 巨 的 備 用 元 件 〇 1 | 透 過 奮 例 以 不 同 技 術 及 / 或 設 計 規 則 製 造 出 來 的 DRAM 1 I 雷 路 中 有 不 同 的 瑕 疵 fcb 例 〇 因 此 即 使 DRAM 電 路 可 能 1 1 例 如 具 有 相 同 的 容 量 也 可 以 在 其 上 提 供 不 同 數 百 的 備 用 1 I 元 件 〇 為 了 易 於 參 閲 * 以 下 將 於 製 造 期 間 應 用 所 發 明 1 I 之 熔 絲 節 省 技 術 的 情 況 稱 為 製 造 - 時 間 的 應 用 〇 1 1 製 造 了 DR AM 電 路 之 後 « 又 可 以 把 熔 絲 節 省 技 術 應 用 在 1 1 -1 0- 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) 經濟部中央標準局員工消費合作社印製 A7 __B7 五、發明説明(9 ) 決定如何將瑕疵元件分配到備用元件以便進行替換。通 常是在製造之後例如在DRAM電路製程中的品管測試期間 才發現這些瑕疵元件。由於所提供備用元件中的熔絲數 目已藉著於製造期間應用所發明之熔絲節省技術而作了 最佳化,重要的是將其熔絲能由適當的共用熔絲之備用 元件群所共用之瑕疵元件作正確地分配。熔絲節省技術 於替換時間内在瑕疵元件上的應用,會有利地確保毎一 個薛疵元件的唯一位址即使在其中某些元件共用熔絲的 情況下也能由備用元件之一唯一地表7K出來。 較佳實施例放說服1 為了作更詳盡的説明,而在利地辨識出在出現相當數 目以上之瑕疵元件時會開始共用它們的高階位元(亦即 更重要的位元)。為了展示這項事實,考慮一個假性DRAM 陣列其中只有十六値主陣列元件。為了唯一地定出這十 六個主陣列元件的位址,必需有四値位址位元(A3, A2 .A1,及A0)。假定吾人於替換時間内發現這十六個主 陣列元件中有五橱是有瑕疵的。第3圔展示的在五個瑕 疵元件及其位址。應注意有三個瑕疵元件〇 〇 0 1 , 0 1 1 0, 及0 111共用著最高階位元A3(其值在此例中為「0」)。 因此,可以辨識出若能於製造期間組合某些備用元件使 得能由共用熔絲代表某些高階位元則能有利地實現在熔 絲而積上的節省。最重要的是,熔絲的共用必須依能節 省最大數目熔絲且不致在替換時間内對可以取代的瑕疵 元件诰成有害衝擊的方式達成,亦即不致産生無法接受 -1 1 - 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) ''裝 訂 M (請先閲讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(、。) 的情況使得其中因為其位址無法由共用熔絲之備用元件 之一表出而有一個不能取代之瑕疵元件。 本發明一個實施例是與習知趨近法相反,假定提供六 個備用元件而能取代DRAM陣列中總共具有十六個主陣列 元件而總數是六痼的瑕疵元件。第4A圖展示的是所提供 的六Μ備用元件,以及通常依-習知技術而提供的熔絲。 應注意此例中,縐共有6 X 4或二十四個位址熔絲(由於 需要四値熔絲才能唯一地定出十六個主陣列元件之一的 位址)。同時還有六個啓動熔絲顯示為啓動熔絲Ε卜Ε 6, 故總共有三十値熔絲。習知設計之第4Β圖展示的是於替 換時間内可以如何習知設計之第4 Α圖中的熔絲以利第3 圖實例中瑕疵元件(亦即其位址為0 0 0 1 , 0 1 1 0 , 1 0 0 1 , 及1U0的瑕疵元件)的替換工作。為求完整,圖中還顯 示了已設定的啓動熔絲。未經設定的熔絲是以符號「X j 表示。 第5圖中所展示的是,根據本發明一個實施例中六値 備用元件以及供相鄰備用元件共用以節省熔絲數目的熔 絲配置。於製造時間内,組合備用元件使得備用元件R 1 -R3會共用最高階的熔絲F3(由熔絲F3與備用元件R4-R6 也共用箸次高階的熔絲F 2。 因此,節省了兩値熔絲F 3 (由於只需一個熔絲F 3以供 備用元件R 1 - R 3之間共用)。此外,也節省了兩脑熔絲F 2 (於R1/R2及R4/R5等群中各節省了一個),故總共節省 了四値熔絲。稍後將參照第8圖而討論所發明用於各群 -12- 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) 、衣 訂 f^ '/{. '1, (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 ___B7___ 五、發明説明(、·) ' 備用元件中執行共用熔絲的運算法則。 第6圖偽第5圖中備用元件的簡略電路圖,其中備用 元件R1-R3之間共用一個熔絲F3。備用元件R1及R2共用一 櫥熔MF2。而備用元件R4及R5刖共用另一個熔絲F2。 第7圖表示的是可以如何設定第5圖中備用元件的熔 絲以利替換第3圖之前述實例中的瑕疵陣列元件(亦即 其位址為〇〇〇〗,0110, 0111,1001,及1110的瑕疵元件) 。由於運轉期間不會使用其啓動熔絲未作設定之備用元 件.故某些斑備用元件R2耦合之共用熔絲是已設定的此 種事奮(例如將F2設定為0)並不會産生誤差。只是無法於 葎轉期間以備用元件R2取代主陣列元件之一。在包含共 用熔絡的情況下,備用元件Rl, R2, R4, R5及R6的熔絲 可完全由第3圖實例中之瑕疵元件(亦即其位址為〇〇〇1, οι 1 ο,〇 m , 1 ο ο 1,及111 〇的瑕疵元件)的個別位址表 出。以下將參照第10圔和第11圖,對這値實例中之替換 技術的應用作更詳盡的討論。 為了進一步展示本趨近?k所具有的彈性,考慮在品管 潮丨試期間只發現三値瑕疵元件0001,0101及1111的狀況 Λ丨比例中,這三橱瑕疵元件可以分別由第5圖中備用元 件m, R3,及R4所取代。以下將參照第10圖和第12圖, 對這個實例中替換技術的應用作更詳盡的討論。 至於另外一橱實例,考慮在替換期間發現瑕疵元件 〇〇〇〇.〇〇〇 1, 0010,0011, 0100,及 1101 的狀況。此例 中,這六個瑕疵元件可以分別由第5圖中備用元件R1, -1 3- 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) --------Ί------、!!-----<lw (請先閲讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(π ) 一 R 2 . R 3 . R 6 , R 4及R 5所取代。以下將參照第1 0圔和第1 3 圖,對這®實例中替換技術的應用作更詳盡的討論。 第8圖展示的是根據本發明中一個概念於製造期間在 共用熔絲目的下用來由備用元件組合成任意的備用元件 群的一般方法,此方法藉著將位元位置計數值X初始化 為0 ,節省下的熔絲總(TS >初始化為0,將Μ設定為所提 供備用元件的數目而由步驟8 0 2開始執行。 於一階眷代(first iteration)中,是處理由其中提供 的所有備用元件所構成的整値群。稍後將證明,會對這 個群作(it era tively)的分割及處理直到這個群(所含備 用元件的數目)小於三為止。緊接著步驟δ 0 2的是,於步 驟804中決定這個群内是否至少具有三個備用元件。若 其中所含備用元件的數目是小於三,則無法在不損害其 替換彈件之下再利用組合備用元件及結合相鄰備用元件 中的熔絲元件而額外地節省所用的熔絲。若這個群中具 有的備用元件是少於三個,則對這個群的處理程序會終 lb於步驟8 2 0。 另一方而若其中所含備用元件的數目是三個或更多, 此方法便開始將具有Μ個元件的備用元件群分割成兩個 群:Α和Β 。於步驟806中,將ΙΝΤ(Μ/2)個備用元件組 合在一起而形成Α群(其中若Μ/2並不是一個整數則運 篁子IHT將M/2的值截取成比下一個更高的整數)。參見 第4A圖,其中有六値備用元件(亦即M=6)。因此,A群 中會含有三個備用元件, R 2 ,及R 3。 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) d 裝 訂 ί^ /(\. (請先閲讀背面之注意事項再填寫本頁) A7 B7 五、發明説明( ) - - 1 1 於 步 驟 8 0 8 中, 將剩餘的備用元件(亦 即 M- INT (M/ 2 ) 1 1 I 個 備 用 ΤΓΐ 件 組 合 在 一 起 而 形 成B 群 〇 參 見 第 4A 圖,B 群 1 1 中 含 有 三 m 備 用 元 件 R4 • R5 , 及 R6 〇 A 和 B群在第 9 請 先 1 1 圖 是 顯 示 於 標 題 為 厂 —^ 階 羅 赞 代」 此 行 底 下 〇 閱 讀 1 背 1 於 步 驟 8 1 〇中, A 群中的各備用元件是共用著於位元 之 1 注 I 位 置 X ( 冃 ,·» j. 刖 的 X = 0 )上的- -些熔絲。 因此第5 圔中的各 意 事 1 備 用 元 件 是 丑 户·、 用 著 熔 絲 F3 ( 位元 位 置 X= 0)。 第9圖中 項 再 ! 也 利 用 備 用 元 件 R 1 « R2 » 及 R3上 各 熔 絲 F3 是 組合在一 起 俱 寫 本 裝 的 堇 實 t-rrT 標 示 出 m 種 共 用 情 形 〇於 步 驟 812 中, 將位元位 頁 1 1 詈 計 敷 倌 X 增 加 1 〇 刖經過這値- -階疊代程序處理之後 1 I 所 節 省 的 熔 絲 總 數 (TS ) 可 以 表為 下 列 方 程 式 0 1 1 TS = TS + INT (M/ 2 -1 ) (方程式 1) 1 訂 其 中 TS代 表 的 是 截 至 巨 .W- 刖 為止 所 節 省 的 熔 絲總數, Μ 1 I 代 表 的 是 \-3_ m 値 锖 代 程 序 中 備 用元 件 的 數 巨 9 而若M/ 2 -1 1 1 並 不 是 一 個 整 數 » 則 蓮 算 子 INT將 M/ 2- 1的值截取成下 1 1 一 m 更 局 的 整 數 〇 由 於 m 個 疊代 程 序 中 備 用 元件的數 巨 k 1 是 -i. 個 9 故 現 在 TS值 為 0 + INT (6/2- 1) 或 者 是 2〇 之 後 » 此 方 法 的 執 行 方 式 是從 步 驟 804開始以A群為 1 l 新 的 第 一 群 備 用 元 件 作 遞 歸 式作 業 > A 群 中 所用的新 Μ 1 1 價 是 設 定 成 等 於 巨 刖 A 群 中 所含 備 用 元 件 的 目的(步驟 1 1 81 4) 目. 其 本 身 的 X 值 是 等 於 目前 的 X 值 〇 由 於目前的 A 1 1 群 中 含 有 三 個 備 用 元 件 (R 1 , R2 , 及R3) 9 故 第一群中 的 1 I Η = =3 (Ί 而 谁 入 二 階 晷 代 時 用 於這 値 第 一 群 之 位元計數 值 1 1 X 是 1 〇 1 I -1 5 - 1 1 1 1 經濟部中央標準局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(κ ) ' I比方法中也依遞歸方式執行以B群當作分開的一群備 用芫件之處理。這個第二群B會將其本身的Μ值設定成 等於Β群中所含備用元件的數目(步驟816)且其本身的 X倌是等於目前的X值。由於目前的Β群中含有三値備 用元件(R4, R5,及R6),故這個第二群中的Μ=3。而在 准人二階《代時用於這個第一群之位元計數值X是1。 之後.同時將第一群和第二群的備用元件、連同它們 铕別的Μ倌及新的位元位置計數值X送回步驟80 4以進 行遞歸作業(步驟818)。第9圖中在標題「一階層代」 底下顯示的是一階疊代程序結束時的結果,並將兩値群 (分別包括備用元件R1/R2/R3及R4/R5/R6)輸入到二 階覉代程序中。這兩値群是當作兩値分開的執行線而進 行處理的亦即兩個分開的程序。 於二階疊代中,可以先處理第一群或是第二群。若先 處理第一群(包括備用元件R1/R2/R3的舊有Α群),其 Μ值為三而不是小於三(步驟8 0 4 )。因此,此方法進行 步驟806而由這値第一群(具有備用元件R1/R2/R3)裝 配出新的A群和新的B群。如先前所述,新的A群會有 TNT(M/2)锢備用元件。因此,新的A群會有兩値備用 元件R1,及R2。而新的B群會有一痼備用元件R3。這種 情形是顯示於第9圖中標題「二階疊代」底下。 於步驟810中,新的A群會共用落在位置X上的熔絲 由於X值為1,故備用元件R1,及R2會共用熔絲F2。 於步驟812中,會將位元計數值X更新為3。完成這値簦 -16- 本紙張尺度適用中國國家標準(CNS ) Α4说格(210Χ297公釐) -----------;'i衣------訂-----γ;咸 /(V·, . ..S (請先閲讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(,〇' 代稈序之後,所節省的熔絲總數(TS)是TS+INT(M/2-l) 或是2+IHT(3/2-l)或是2 + 1或是3。第9圖中在標題「 二階轉代i底下顯示的是對包括備用元件R1及R2的群施 行二階瘅代程序結束時的結果。 將根據步驟814, 816,及818,而對這兩個新的群(分 別包栝備用元件ΙΠ / R 2及R 3 )進行遞歸式處理。不過, 由於這兩餹群在三階驀代程序開始時都沒有通過步驟804 的測試(亦即至少具有三個備用元件),故對這兩個新的 群的處理程序會終止於步驟820。 於二階番代中,將二階疊代的第二群(包括三値備用 元件R 4 , R 5 ,及R 6 )當作分開的執行線而進行類似的處 理。對這個具有三値備用元件R4, R5,及R6之第二群的 曦理,在將這櫥第二群分成兩値群且備用元件R4和R5共 用熔絲F 2 (參見第9圖)時也産生又節省了 一値熔絲的結 果。所節省的熔絲總數(TS)是TS+IHT(M/2-l)或是3+INT (3 / 2 - 1)或是3 + 1或是4。 將稂據步驟814, 816,及8 18,而分別處理從包括備 用元件R4/R5/R6分出的這兩値新的群。參照第9圖中 的實例,這兩値新的群分別包括備用元件R 4 / R 5及R 6。 不過,由於這兩痼群在三階叠代程序開始時都没有通過 步驟804的測試(亦即至少具有三個備用元件),故對這兩 锢新的群的處理程序會終止於步驟820。 由第5圖及第9圖中可知,第8圖之方法的應用産生 了可節省四個熔絲(兩個熔絲F3以及兩個熔絲F2)的結果 -17- 本紙張尺度適用中國國家標準(CNS ) Α4規格(210X297公釐) I 11 n·,., 4 裝 n Ύ 嫌 ./' :-··· (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(4) ' … ,這橱信是表為二階疊代程序結束時的TS值。將第8圔 之一般方法應用於具有任意數目之備用元件之任意備用 元件群的情形則留給讀者當作練習的作業。 一曰.製成具有共用熔絲之備用元件的DRAM 1C,可以 測試,此D R A Μ I C以決定主陣列元件中有多少個瑕疵元件 目究竟是那些元件具有瑕疵。然後將這些有瑕疵的陣列 元件取代成共用熔絲之備用元件。根據本發明中另一値 概念.太發明會提供一種方法以便將瑕疵元件分配到共 用熔絲之備用元件上。由於現在有某些備用元件會共用 熔絲.重要的是將瑕疵元件群分配到正確的備用元件群 上使之不致損害其替換彈性,亦即不致於無法取代一個 或更多的薛疵元件。 第10圖展示的是根據本發明中一個概念於替換期間在 後續將毎一個瑕疵元件分配到共用熔絲之備用元件中的 滴當備用元件上之目的下用來組合瑕疵元件的方法。此 方法藉箸加入足夠數目之具有唯一位址的偽造瑕疵元件 (亦即與真實瑕疵元件位址不同的非重複性位址)使得瑕 疵元件的總數(真實的加上偽造的)等於所提供備用元件 的數目而由步驟1 0 0 1開始執行。通常,偽造的瑕疵元件 是設有旗標使得在將它們分配到備用元件時,這些備用 元件上的相關啓動熔絲是没有設定的(由於這些瑕疵元 件是偽造的R是為了作分配而加進去的而不是要作真實 的替換之用以下將要展示的是,吾人是依不明顯的 新奇方式加入偽造瑕疵元件而無論在製造期間發現多少 _ 1 8 一 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) 裝 訂 (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(Γ _ 橱瑕疵元件(上至所提供備用元件的總數)都允許作出正 確的分配。 於步驟1 0 0 2中,將位元位置計數值X初始化為0並將 Μ設定為所提供備用元件的數目。在執行步驟1002之前 .若必要的話可依位址分類篩選這些備用元件(例如依 遞增之次序)。 除一階#代中,是處理由其中發現到的所有瑕疵元件 所構成的整個群。稍後將要證明,吾人會對這個群作疊 代式的分割及處理直到這個群(所含瑕疵元件的數目)小 於三為出。緊接著步驟1 0 0 2的是,於步驟1 0 0 4中決定這 値群内是否至少具有三個瑕疵元件。若其中所含瑕疵元 件的數目是小於三,則對這値群的處理程序會終止於步 驟1 0 2 0,且可能立即開始分配工作。稍後將會討論在處 理之後的分配工作。 另一方而若其中所含瑕疵元件的數目是三個或更多, 此方法便開始將具有Μ個元件的瑕疵元件群分割成兩個 群:Α和Β 。於步驟1006中,將ΙΝΤ(Μ/2)値在位元位 置X上共用一個共同位元值(不是0就是1)的瑕疵元件 組合在一起而形成Α群(其中若Μ/2並不是一痼整數, 刖葎算子INT將M/2的值截取成比下一値更高的整數)。 參見第3圖中的實例,其中有五個瑕疵元件(亦即M=5) 。因此,會加入一痼具有唯一位址的偽造瑕疵元件使得 瑕疵元件的總數(真實的以及偽造的)等於所提供之備用 元件的數目。如第11圖所示,這個偽造瑕疵元件是顯示 -19- 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) --------〆裝------訂-----X,¥ (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明(d ) 成偽诰瑕疵元件D2(其位址為0010)。這個偽造瑕疵元件 b势有旗標(第11圖中標示為*號)以確保在將它們分配 到備用元件之後,這些備用元件上的相關啓動熔絲是没 有設定的。於第11圖的實施例中,也在加入偽造瑕疵元 件之後對瑕疵元件的原始群依遞增次序進行分類筛選。 A群會由三個·瑕疵元件Dl, D2,及D3組成。這値群在 第1 1圖是頴示於檫題為「一階疊代」此行底下。由於瑕 疵元件D 1 , D 2 ,及D 3的熔絲F 3都有相同的位元值「0 j , 故痦種組合稈序是TH確的。 於步驟1 0 0 8中,將剩餘的瑕疵元件(亦即Μ - I N T ( Μ / 2 ) 個瑕疵元件組合在一起而形成Β群。參見第3圖中的實 例,Β群中會含有三痼瑕疵元件D4, D5,及D6。這個群 在第1 1.圖也是顯示於標題為「一階疊代」此行底一。 於步驟10 10中,Α群中的各瑕疵元件是共用著於位元 位詈X (目前的X=0)上的一些熔絲。因此,第11圖中的 抿疵元件Dl, D2,及D3是共用箸熔絲F3(位元位置X=0) 於步驟1 0 1 2中,將位元位置計數值X増加1。 之後.此方法執行方式是從步驟1 0 0 4開始以A 群為新 的第一群備用元件而作遞歸式處理,亦即現在是以A群 本身當作分開之執行線進行處理。為了方便遞歸式作業 ,A群中所用的新Μ值是設定成等於目前A群中所含瑕疵 元件的數目(步驟ΐ 0 1 4 )且其本身的X值是等於目前的X 倌。由於目前的Α群中含有三個瑕疵元件(Dl, D2,及 D3),故第一群中的M=3。而在進入二階疊代時用於這 -20- 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) ; γ裝 訂 Ϊ ..V (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 ___B7___ 五、發明説明(―勹…广 個第一群之位元計數倌X是1。 眈方法中也依遞歸方式執行以B群當作分開的一群瑕 疵元件之處理。這個第二群B會將其本身的Μ值設定成 等於Β群中所含接線元件的數目(步驟1016)且其本身的 X倌是等於目前的X倌。由於目前的Β群中含有三値備 用元件(D4,D5,及D6,故這値第二群中的Μ=3。而在 谁λ二階羁代時用於這値第一群之位元計數值X是1。 之後.同時將第一群和第二群的備用元件、連同它們 個別的Μ信及新的位元位置計數值X送回步驟1 0 〇 4以進 行镢歸式作業(步驟1018)。第11圖中在標題「一階疊代」 底下顯示的是一階晷代程序結束時的結果,並將兩値群 (分別包栝瑕疵元件D1/D2/D3及D4/D5/D6)輸入到二 階眷代程序中。這兩個群是當作兩個分開的執行線而進 行處理的,亦即兩値分開的程序。~ 於二階叠代中,可以先處理第一群或是第二群,若先 處理第一群(包括瑕疵元件D1/D2/D3的舊有Α群),其 Μ值為三而不是小於三(步驟1 0 0 4 )。因此,此方法進行 步驟100(5而由這個第一群(具有瑕疵元件D1/D2/D3)裝 配出新的Α群和新的Β群。如先前所述,新的Α群會有 TNT(M/2)値瑕疵元件。由於包括有D1/D2/D3等瑕疵 元件群中的瑕疵元件D 1和D 2於目前的位元位置計數值X (目前的X=l)上共用相同的位元值「0」,故將瑕疵元 件Μ和1)2組合在一起而形成新的A群。因此,新的A群 會有兩個瑕疵元件D1和D2。而新的B群會有一艏備用元 -21 - 本紙張尺度適用中國國家標隼(CNS ) A4規格(210X297公釐) (請先閱讀背面之注意事項再填寫本頁) 訂 B7__ 五、發明説明(“) ~' "" 伴D 3。這種情形是顯示於第1 1圖中標題「二階昼代」底 下。 於步驟101 〇中,新的A群會共用落在位置X上的熔絲 。由於現在X價為1,故瑕疵元件D1和D2會共用熔絲F2。 由於01和1)2¾兩橱瑕疵元件都在位元位置χ=ΐ上共用共 同的位元倌「〇 I ,故這種組合程序是正確的。於步驟 1012中.會將位元計數值X更新為3。第11圖中在標題 「二階#代!底下顯示的是對包括瑕疵元件D1和D2的群 施行二階眷代稈序結束時的結果。 將根據步驟1〇 14, 1〇16,及1018,而對這兩値新的群1T »1 'A7 B7 Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs 5. Description of the Invention (4-)-1 1 Defective main array element 9 However, there are still some shortcomings 〇 Example 1 1 I ♦ Increasing the capacity of the DR AM chip requires placing a larger hundreds of main arrays 11 1 billion body components and / or spare components into a given size wafer. Do not read 1 1 if * each fuse is in the DRA Occupy __. 値 on the crystalline crystal 値 inconspicuous face 1 back 1 makes the large number of giant fuses required in the conventional approach approach disadvantageously required 1 I larger die (Die) size to accommodate All fuses. '-' Italian 1 For display purposes i If there are 2 1 値 components in the array, each-'spare item • 1 ii components must have η cabinet address fuses 〇 each spare component must also be written Start the fuse 9 once to indicate whether page 1 1 should be used during operation as discussed above. Page 1 1 This spare element should be used. Therefore i in the conventional design each spare 1 I 7Π The total number of fuses must be η + 1 〇If every 1-arrays are provided with 1 spare components (with replacement of m defective components in the main array) 1 The total number of fuses required for ordering is m * (η + 1) 〇 It is worth knowing The note of the conventional technology 1 I is that if the total number of fuses can be reduced, the mold size can be reduced 1 1, the size of the DR AM crystal can be reduced) and / or the density of the DRA M 0 1 1 ) Improved technology is expected to provide spare components for DR AM arrays. Advantageously reduces the area required for its implementation * so that it can 1 place a larger number of main array elements and / or 1 1 spare elements in a given size wafer 〇1 | Invention_ Μ-1 1 in one In the embodiment of the cabinet, the present invention relates to a method for making a number of first spare parts in a memory device. 1 | The method of making many first spare parts to reduce the number of address fuses used is hundreds. Number of spare components 1 1 Large is at least one. If many 6 Number of spare components in the first spare component 1 1 1 1 1 This paper size applies to China National Standard (CNS) A4 (210X297 mm) kl B7 Central Ministry of Economic Affairs Standard Bureau employee consumer cooperative printing 5. Invention description (r) \ 1 is at least three. This method includes the first spare component group formed by many first spare elements 1 1 I pieces. Steps to optimize the site fuse. This 1 1 optimization step also includes the formation of a second spare from many first spare components. Please use the component group first. 1 The second spare component group and the first spare component. The group ik 1 and the back 1 are η-rejected. This optimization step also includes providing a stack of first fuses 1¾ of 1 fuse Note | for the use of the spare components in the first spare component group. The first idea 1 forms the most local address fuse of each spare element in the first spare element group. Item 1 / In another embodiment, the present invention relates to a memory array, and there are many The __ * spare element of the common fuse replaces the defective component of page S. 1 1 in the memory array. The memory K array contains a first fuse and the first by many I 1 common fuses. Prepare The first spare component group composed of components 〇1 1 Each of the first spare component groups has a shared first fuse as their highest order 1 address fuse. The billion-body array additionally contains a plurality of shared fuses. The first I-second spare component group constituted by the second spare component group. The second spare I 1 component group and the first spare component group are mutually exclusive. In another embodiment, the present invention relates to a computer. -Implementation method,. 丄-, / '' is used to replace many of the first defective components in the memory array with a number of spare components in the first spare component of the fuse. 0 kinds of computers-Implementation 1 The method includes determining whether the number of defective components in many first defective components is 1 I less than the number of spare components in first spare components in many common fuses. 1 1 Step 〇 If many first defects Hundreds of defective components in the defective component are smaller than the number of spare components in the first spare component of many 1 I shared fusions. Then this electrical 1 1 brain-implementation method is also included in many defective components. 1 1 Pseudo-defective components in order to transfer 1 of the defective components among many 7 defective components 1 1 1 1 1 This paper size applies to China National Standard (CNS) A4 specifications (210X297 mm) A7 B7 Employees of the Central Bureau of Standards, Ministry of Economic Affairs Printed by the cooperative V. Description of the invention (b)--1 1 Hundreds of 1 1 I number of spare parts added to the first spare part of many shared fuses 0 1 1 This computer-implementation method also includes determining many -RS in defective components, please first 1 1 whether there are at least three rounds of defective components. If many of the first defective components are read, 1 there are at least three hundreds of defective components in the back 1-this kind of computer Method 1¾ of 1 Note | Contains the steps of performing the 1st-defective component group idea 1 from a number of defective components by replacing the individual components of the many first defective components with a number of terms; j more first spares Individual components in the component. 0 Each fill in the first defective component group. 1 Defective components share the same bit value page at the highest-order address bit position. «_〆II η m substitution steps additionally include The step of forming the first II defective component group by many first defective components. The second defective component group and the first defective component group II are mutually interactive. This replacement step also includes the order of the first defective component group I. Individual components are assigned to individual components in the first spare component group so that the steps of I | In the following details, reference is made to the attached 圔 II for a more detailed description. 'Ϊ́ Schematic description I The present invention is shown as an example, but it is not limited to this> The same reference in the drawings of the attached chart II The code marks similar components. II. Brief description of the II diagram. Figure 1 shows the DRA MU unit with a m main II array that has been simplified for the sake of m. 0 1 1 1 Figure 2 shows The spare circuit diagram used in the conventional design refers to the replacement address of 1 1 spare components. 0 1 I 8 1 1 1 1 This paper size applies to China National Standard (CNS) A4 specification (210 X 297 mm) A7 B7 Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs. 5. Description of the invention (7) 1 1 1 Figure 3 For the convenience of discussing the possibility of sharing fuses The display 1 1 1 shows the five defective components for interpretation and their addresses. 0 1 I The 4 A HI display is provided in the conventional design-each one has four. Please read the I address fuse first. All fuses when spare parts are read. 1 I Figure 4B shows how the ficus in Figure 1 in Figure 4A can be set during the replacement time to facilitate the replacement of the defective JQ part in the example in Figure 3. Event 1 Figure 5 shows the shared item according to the present invention—an embodiment of the invention. 1 ** ν 1 \. The spare parts with fuses and the spare parts shared by adjacent spare parts to save the number of fuses. Page 1冃 Fuse configuration 0V ✓ 1 Simplified circuit of spare parts in Figure 6 and 55 圔 1 | Figure 7 shows how to set the 5th The spare part 1 in the figure | fuse to replace the 3rd part of the picture .. »» Λ. Defective array element in the example described below. The order shown in Fig. 8 shows a concept according to the present invention during manufacturing. 1 I in The method of combining spare parts for the common fuse giant is shown in Figure 1. Figure 9 shows the method of applying Figure 8 in order to optimize the fuse number provided in the conventional design of Figure 4A. 〇 丄 W Figure 10% shows the use of the concept of the present invention in the replacement period 1 in the subsequent allocation of each defective component to the spare parts of the common fuse 1 of the appropriate spare parts of 100 Method for combining defective components 〇1 I Fig. 11 shows the application of the method of Fig. 10 in Fig. 3 Example 1 I 〇 1 1 Fig. 12 Λ shows the application of Fig. 10 square m is used on a group of defective components whose 1 I addresses are 000 1 0101 > and 1111 respectively. 1 1 Fig. 1 3 Fig.> shows the application of the method of Fig. 9 10 to the group "— ▲ 値 Its 1 1 1 1 1 This paper size is applicable to Chinese National Standards (CNS) A4 specifications (210X297 mm) A7 B7 Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economy and Economy 5. Description of the invention (S) 1 1 Yes 0000 »000 1 9 00 10 9 0011 9 01 00 9 and 110 1 of 1 1 I on the defective component 0 1 1 Detailed description of the invention please A, 1 1 Now referring to the examples and The drawings are used to illustrate this reading 1 back 1 invention 〇 in the following description »set up a lot of details to provide 1 I understanding of the τη of the invention 〇 but for those who are familiar with conventional technology -Λ-. ≫ intended 1 Obviously this The invention can be carried out under some or all of the appended patent application scopes under the Li-. Λ 'item. In other situations, in order not to impose unnecessary obstacles on the theory of the invention Λ7Ϊ m > This is not ~ r ^ ~ I will give a detailed description of the processing steps that I am familiar with Page 1 1 / or the structure 〇 1 I Generally speaking> The invented fuse saving technology can be used at the same time as the package 1 1 with DR AM s Synchronous DRAMs (SDRAMs) and other dynamic random access memory devices (R AM S), such as the manufacturing period of billion-volume volume circuit (ICs), and the replacement period of the main array element of the defect 1 I other 1C s contains logic crystals such as the gate programmable array 1 (FPG As) or other logic 1 1 crystals. 〇 During the manufacturing of DR AM, fuse saving technology can be used to determine how The fuses are shared between the spare parts provided. Generally, there are many different types of spare components in different types of DR AM circuits. 01 | DRAM manufactured through different examples with different technologies and / or design rules. 1 I have different defects in the lightning circuit. The DRAM circuit may 1 1 for example have the same capacity or it can provide hundreds of different spare 1 I components on it. For easy reference * the case where the invented fuse saving technology of 1 I will be applied during manufacturing is called manufacturing -Application of time 〇1 1 After manufacturing the DR AM circuit «Fuse saving technology can be applied to 1 1 -1 0- 1 1 1 This paper size is applicable to China National Standard (CNS) A4 specification (210X297 mm) Economy Printed by the Consumer Standards Cooperative of the Ministry of Standards, Ministry of Standards A7 __B7 V. Description of Invention (9) Decide how to assign defective components to spare components for replacement. These defective components are usually found after manufacturing, such as during quality control testing in DRAM circuit manufacturing processes. Since the number of fuses in the spare parts provided has been optimized by applying the invented fuse saving technology during manufacturing, it is important that their fuses can be replaced by an appropriate spare part group that shares a fuse. Common defective components are correctly distributed. The application of fuse saving technology on defective components during replacement time will advantageously ensure that the unique address of a defective component can be uniquely derived from one of the spare components even if some of them share a fuse. . The preferred embodiment persuades 1 to make a more detailed explanation, and to identify the higher-order bits (that is, the more important bits) that will start to share when there are a considerable number of defective components. To demonstrate this fact, consider a pseudo DRAM array with only sixteen 値 main array elements. In order to uniquely determine the addresses of the sixteen main array elements, there must be four address bits (A3, A2, A1, and A0). Suppose we found that five of the sixteen main array elements were defective during the replacement time. The third part shows five defective components and their addresses. It should be noted that there are three defective components 〇 0 1, 0 1 1 0, and 0 111 sharing the highest order bit A3 (its value is "0" in this example). Therefore, it can be recognized that if some spare components can be combined during manufacturing so that certain higher-order bits can be represented by a common fuse, the savings in fuse accumulation can be advantageously achieved. Most importantly, the sharing of fuses must be achieved in such a way that the maximum number of fuses can be saved without causing harmful impact on the replaceable defective components within the replacement time, that is, no unacceptable -1 1-this paper size Applicable to China National Standard (CNS) A4 specification (210X297 mm) '' Binding M (Please read the notes on the back before filling this page) Printed by the Consumer Standards Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Invention Description (,. ), Which causes an irreplaceable defective component because its address cannot be expressed by one of the spare components of the common fuse. An embodiment of the present invention is the opposite of the conventional approach, assuming that six spare elements are provided instead of the defective elements with a total of sixteen main array elements in the DRAM array and a total of six frames. Figure 4A shows the six M spare components provided, as well as the fuses typically provided by conventional techniques. It should be noted that in this example, there are 6 x 4 or twenty-four address fuses (since four pin fuses are required to uniquely determine the address of one of the sixteen main array elements). At the same time, there are six start-up fuses shown as start-up fuse eb 6, so there are a total of thirty fuses. Figure 4B of the conventional design shows how the fuse in Figure 4A can be learned during the replacement time to facilitate the defective component in the example in Figure 3 (that is, its address is 0 0 0 1, 0 1 1 0, 1 0 0 1, and 1U0 defective parts). For completeness, the figure also shows the start-up fuses that have been set. Unset fuses are indicated by the symbol "X j." Figure 5 shows a six-element spare element and a fuse configuration for adjacent fuse elements to save the number of fuses according to an embodiment of the present invention. During the manufacturing time, the spare components are combined so that the spare components R 1 -R3 share the highest order fuse F3 (the fuses F3 and the spare components R4-R6 also share the next higher-order fuse F 2. Therefore, savings are achieved Two fuses F 3 (because only one fuse F 3 is needed for sharing between spare components R 1-R 3). In addition, two brain fuses F 2 (for R1 / R2 and R4 / R5 etc.) One in each group was saved), so a total of four fuses were saved. The invention for each group will be discussed later with reference to Figure 8. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297) Li), clothes binding f ^ '/ {.' 1, (Please read the notes on the back before filling this page) Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs A7 ___B7___ V. Description of the invention (, ·) '' Spare parts The algorithm of the common fuse is executed in Fig. 6. Simplified circuit of the spare element in Fig. 6 and Fig. 5. Among them, one fuse F3 is shared between the spare components R1-R3. The spare components R1 and R2 share a fuse MF2. The spare components R4 and R5 刖 share another fuse F2. Figure 7 shows how to set the first fuse The fuse of the spare component in Figure 5 is to replace the defective array component in the previous example in Figure 3 (that is, the defective component whose address is 〇〇〇, 0110, 0111, 1001, and 1110). Due to the operation period It will not use spare components whose startup fuses are not set. Therefore, some common fuses coupled with some spot spare components R2 are already set up (such as setting F2 to 0) and will not cause errors. It is just impossible During the turnaround period, one of the main array elements is replaced by a spare element R2. In the case of a shared fuse, the fuses of the spare elements R1, R2, R4, R5 and R6 can be completely replaced by the defective element in the example in Figure 3 ( That is, the individual addresses of defective components whose addresses are 00001, οι 1 ο, 〇m, 1 ο ο 1, and 111 〇) are listed. The following will refer to Figures 10 圔 and 11 for this.値 The application of replacement technology in the example is discussed in more detail. The elasticity of this approach? K. Considering that only three defective parts 0001, 0101, and 1111 are found during the quality control tide, the three defective parts can be replaced by spare parts in Figure 5 Replaced by m, R3, and R4. The application of the replacement technology in this example will be discussed in more detail with reference to Figures 10 and 12 below. As for another example, consider finding defective components during replacement. 〇〇〇 〇〇〇〇〇〇1, 0010, 0011, 0100, and 1101. In this example, these six defective components can be replaced by the spare components R1, -1 in Figure 5 3- This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -------- Ί ------, !! ----- < lw (Please read the notes on the back before filling this page) Printed by the Consumer Cooperatives of the Central Bureau of Standards of the Ministry of Economic Affairs A7 B7 V. Invention Description (π)-R 2. R 3. R 6, R 4 and R 5 Was replaced. The application of the replacement technology in this example will be discussed in more detail with reference to Figures 10 圔 and 13 below. FIG. 8 shows a general method for combining spare components into an arbitrary spare component group for the purpose of a common fuse during manufacturing according to a concept of the present invention. This method initializes a bit position count value X to 0, the total number of saved fuses (TS > is initialized to 0, M is set to the number of spare components provided and is executed from step 8 02. In the first iteration, the processing is performed by A tidy group made up of all the spare elements provided. It will be proven later that this group is divided and processed until the group (the number of spare elements contained) is less than three. This is followed by step δ 02: It is determined in step 804 whether there are at least three spare components in the group. If the number of spare components contained therein is less than three, it is not possible to reuse the combined spare components without damaging its replacement parts and Combine the fuse elements in adjacent spare elements to save additional fuses. If there are less than three spare elements in this group, the processing procedure for this group will end in steps 8 2 0. On the other hand, if the number of spare components contained therein is three or more, the method starts to divide a spare component group with M components into two groups: A and B. In step 806, Combine INT (M / 2) spare components to form A group (where M / 2 is not an integer, then IHT truncates the value of M / 2 to a higher integer than the next one). See also Figure 4A, there are six spare parts (that is, M = 6). Therefore, group A will contain three spare parts, R 2 and R 3. This paper size applies the Chinese National Standard (CNS) A4 specification ( 210X297 mm) d Binding ^^ ((. Please read the notes on the back before filling out this page) A7 B7 V. Description of the invention ()--1 1 In step 8 0 8, place the remaining spare components ( That is, M-INT (M / 2) 1 1 I spare ΓΓ pieces are combined to form group B. See Figure 4A. Group B 1 1 contains three m spare elements R4 • R5, and R6 〇 A and B Groups in the 9th please first 1 1 The picture is shown in the heading Factory— ^ Order Rozande At the bottom of this line, read 1 back 1 In step 8 1 0, each spare element in group A is shared at bit 1 Note I position X (X, 0 of j, · »j. 刖) --Some fuses. Therefore, each of the items in the 5th item 1 The spare element is ugly, and it uses fuse F3 (bit position X = 0). The item in Figure 9 is again! Also use spare parts R 1 «R2» and fuses F3 on R3 are combined together to write m t-rrT to indicate m common situations. In step 812, set the bit The bit page 1 1 is calculated by adding 1 〇 刖 After this--step iterative process 1 1 The total number of fuses saved (TS) can be expressed as the following equation 0 1 1 TS = TS + INT ( M / 2 -1) (Equation 1) 1 where TS represents the total number of fuses saved up to mega.W- 刖, and M 1 I represents the number of spare components in the \ -3_ m program. Giant 9 and if M / 2 -1 1 1 is not an integer », then the lotus operator INT truncates the value of M / 2- 1 to the next 1 1-m more rounded integer, because it is reserved in m iterative programs The number of components k 1 is -i. A 9 so the TS value is now 0 + INT (6 / 2- 1) or 2 0 »The method of this method is to start from step 804 with A group as 1 l new First group Use components for recursive operations> The new M 1 1 price used in group A is set equal to the purpose (step 1 1 81 4) of the spare components included in giant A group. Its own X value is equal to the current Since the current A 1 1 group contains three spare elements (R 1, R 2, and R 3) 9, 1 I in the first group I = = 3 (Ί, and who is The bit count of the first group is 1 1 X is 1 〇 1 I -1 5-1 1 1 1 Printed on the paper by the Consumers' Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs, the paper size applies the Chinese National Standard (CNS) A4 specification (210X297 (Mm) Printed by the Consumers 'Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Description of Invention (κ)' I ratio method also uses the B group as a separate group of spare files in a recursive manner. This second group B will set its own M value equal to the number of spare components contained in the B group (step 816) and its own X 本身 will be equal to the current X value. As the current B group contains three tritium spare elements (R4, R5, and R6), M = 3 in this second group. The bit count value X used for this first group in the second generation of quasi-human generation is 1. After that, the spare elements of the first group and the second group are sent back to step 804 together with their Μ 铕 and the new bit position count value X for recursive operation (step 818). Figure 9 shows the results at the end of the first-order iterative process under the heading "First-order generation", and the two groups (including the spare components R1 / R2 / R3 and R4 / R5 / R6) are input to the second-order In the generation program. These two groups are treated as two separate execution lines, that is, two separate programs. In the second-order iteration, the first group or the second group can be processed first. If the first group (the old A group including the spare elements R1 / R2 / R3) is processed first, its M value is three instead of less than three (step 804). Therefore, the method proceeds to step 806 and a new A group and a new B group are assembled from the first group (with spare elements R1 / R2 / R3). As mentioned earlier, the new A group will have TNT (M / 2) 锢 spares. Therefore, the new A group will have two spare elements R1, and R2. The new B group will have a spare element R3. This situation is shown under the title "Second Order Iteration" in Figure 9. In step 810, the new A group will share the fuses falling at the position X. Since the value of X is 1, the spare components R1 and R2 will share the fuse F2. In step 812, the bit count value X is updated to 3. Complete this 値 簦 -16- This paper size applies the Chinese National Standard (CNS) Α4 grid (210 × 297 mm) -----------; 'i clothing ------ order --- --γ; Xian / (V ·,. ..S (Please read the notes on the back before filling out this page) Printed by A7 B7, Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economy After that, the total number of fuses saved (TS) is TS + INT (M / 2-l) or 2 + IHT (3 / 2-1) or 2 + 1 or 3. Figure 9 shows the title " Shown below the second-order generation i is the result of the end of the second-order generation process on the group including the spare elements R1 and R2. The two new groups (respectively reserve spares) will be based on steps 814, 816, and 818. Components I Π / R 2 and R 3) are processed recursively. However, since neither unitary group passed the test of step 804 at the beginning of the third-order generational program (that is, it has at least three spare components), so The processing procedure for the two new groups will end at step 820. In the second-order generation, the second group of the second-order iteration (including the three-unit spare elements R 4, R 5, and R 6) is regarded as a separate execution line. And proceed similarly For this second group of three groups of spare elements R4, R5, and R6, the second group of this cabinet is divided into two groups and the spare elements R4 and R5 share the fuse F 2 (see Figure 9) It also results in saving another stack of fuses. The total number of fuses saved (TS) is TS + IHT (M / 2-l) or 3 + INT (3/2-1) or 3 + 1 Or 4. According to steps 814, 816, and 8 18, the two new groups separated from the spare components R4 / R5 / R6 are processed separately. Referring to the example in FIG. 9, the two new groups The groups include spare components R 4 / R 5 and R 6 respectively. However, since neither unitary group passed the test of step 804 at the beginning of the third-order iteration process (that is, it has at least three spare components), The processing procedure for these two new clusters will end at step 820. As can be seen from Figures 5 and 9, the application of the method of Figure 8 results in the saving of four fuses (two fuses F3 and two Result of fuse F2) -17- This paper size applies Chinese National Standard (CNS) A4 specification (210X297mm) I 11 n ·,., 4 Pack n Ύ suspect./ ':-·· (Please read first Note on the back (Write this page) A7 B7 printed by the Consumer Cooperatives of the Central Bureau of Standards of the Ministry of Economic Affairs 5. Description of the Invention (4) '… This letter is the TS value at the end of the second-order iterative procedure. Apply the general method of Section 8 In the case of any spare component group having any number of spare components, it is left as an exercise for the reader. It is said that the DRAM 1C with spare components for common fuses can be tested. This DR A MIC can determine how many defective components there are in the main array component, and whether those components have defects. These defective array elements are then replaced with spare elements that share a fuse. According to another concept in the present invention, the invention will provide a method for allocating defective components to a spare component of a common fuse. Because some spare components now share fuses, it is important to assign defective component groups to the correct spare component group so as not to impair their replacement flexibility, that is, they cannot replace one or more defective components. Fig. 10 shows a method for combining defective components during replacement in accordance with a concept of the present invention for the subsequent allocation of one defective component to a spare component of a common fuse. This method adds a sufficient number of forged defective components with unique addresses (that is, non-repeating addresses that are different from the actual defective component addresses) so that the total number of defective components (real plus forged) is equal to the spare provided. The number of components is executed from step 1001. Usually, the fake defective components are flagged so that when they are allocated to the spare components, the relevant startup fuses on these spare components are not set (because these defective components are fake and R is added for distribution) It is not to be used for real replacement. What will be shown below is that we add forged defective components in a novel way that is not obvious, no matter how many are found during manufacturing. 1 8 A paper size applies the Chinese National Standard (CNS) A4 Specifications (210X297 mm) Binding (please read the precautions on the back before filling this page) Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Description of the invention (Γ _ cabinet defective components (up to the spare components provided The total number is allowed to make the correct allocation. In step 10 02, the bit position count value X is initialized to 0 and M is set to the number of spare components provided. Before performing step 1002, if necessary, follow the steps Address classification to filter these spare components (for example, in ascending order). In the first-generation # generation, all defects found in it are processed. The entire group of components. We will prove later that we will iteratively divide and process this group until the group (the number of defective components) is less than three. The next step is 1 0 0 2 is In step 10 0, it is determined whether there are at least three defective components in the group. If the number of defective components contained in the group is less than three, the processing procedure for the group is terminated in step 1 0 2 0, And the distribution work may start immediately. The distribution work after processing will be discussed later. On the other hand, if the number of defective components contained therein is three or more, the method will start with a defective component group with M components Divided into two groups: A and B. In step 1006, the defective elements of INT (M / 2) 不是 sharing a common bit value (either 0 or 1) at bit position X are combined to form A Group (where M / 2 is not a unitary integer, unitary operator INT truncates the value of M / 2 to a higher integer than the next unitary.) See the example in Figure 3, where there are five defective elements (That is, M = 5). Therefore, a One-site forged defective components makes the total number of defective components (real and forged) equal to the number of spare components provided. As shown in Figure 11, this forged defective component is displayed-19- This paper size applies to China Standard (CNS) A4 specification (210X297mm) -------- Outfit -------- Order ----- X, ¥ (Please read the precautions on the back before filling this page) Economy Printed by the Consumer Standards Cooperative of the Ministry of Standards of the People's Republic of China A7 B7 V. Description of the invention (d) A fake defective component D2 (its address is 0010). This fake defective component b has a flag (marked as * in Figure 11) ) To ensure that after assigning them to spare components, the associated startup fuses on those spare components are not set. In the embodiment of FIG. 11, the original group of defective components is classified and screened in ascending order after the forged defective components are added. Group A will consist of three defective elements Dl, D2, and D3. This group is shown in Figure 11 below the line entitled "First Order Iteration". Since the defective elements D 1, D 2, and F 3 of D 3 have the same bit value “0 j, the sequence of each combination is TH correct. In step 10 0, the remaining Defective components (that is, M-INT (Μ / 2) defective components are combined to form a B group. See the example in Figure 3. The B group contains triplet defective components D4, D5, and D6. This group In Figure 11.1, the figure is also shown at the bottom of the line titled "First Order Iteration". In steps 10 and 10, each defective component in the A group is shared by the bit position 詈 X (current X = 0 ). Therefore, the defective components D1, D2, and D3 in Fig. 11 are common fused fuses F3 (bit position X = 0). In step 10, the bit position is counted. The value X is incremented by 1. After that, the method is executed recursively from step 1 0 4 with A group as the new first group spare component, that is, now with A group itself as a separate execution line In order to facilitate recursive operations, the new M value used in group A is set equal to the number of defective components currently contained in group A (step ΐ 0 1 4) and its own X value is equal to The former X 倌. Since the current A group contains three defective elements (Dl, D2, and D3), M = 3 in the first group. It is used for this -20-paper when entering the second-order iteration. Standards are applicable to China National Standard (CNS) A4 specifications (210X297 mm); γ binding Ϊ..V (Please read the notes on the back before filling this page) Printed by the Consumers Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 ___B7___ V. Invention Explanation (― 勹… the first group of bit counts 倌 X is 1. 眈 The method also uses the group B as a separate group of defective components in a recursive manner. This second group B will The value of M is set to be equal to the number of wiring components included in the B group (step 1016) and its own X 倌 is equal to the current X 倌. Since the current B group contains three 値 spare components (D4, D5, and D6, Therefore, M = 3 in the second group. When the lambda second-order generation is used, the bit count value X of the first group is 1. After that, the spare elements of the first group and the second group are simultaneously used. , Together with their individual M letter and the new bit position count value X are sent back to step 104 for homeopathic operation Step 1018). Figure 11 shows the results at the end of the first-order generation process under the heading "First-order iteration." The two groups (including defective elements D1 / D2 / D3 and D4 / D5, respectively) / D6) into the second-order dependent program. These two groups are processed as two separate execution lines, that is, two separate programs. ~ In the second-order iteration, the first group can be processed first Or the second group, if the first group (including the old A group of defective components D1 / D2 / D3) is processed first, its M value is three instead of less than three (step 1 0 0 4). Therefore, the method proceeds to step 100 (5 and a new A group and a new B group are assembled from this first group (with defective elements D1 / D2 / D3). As described earlier, the new A group will have TNT ( M / 2) 値 Defective components. Defective components D 1 and D 2 in the defective component group including D1 / D2 / D3 share the same bit on the current bit position count value X (current X = 1) The element value is "0", so the defective elements M and 1) 2 are combined to form a new A group. Therefore, the new A group will have two defective components D1 and D2. And the new B group will have a spare yuan-21-This paper size applies to the Chinese National Standard (CNS) A4 size (210X297 mm) (Please read the precautions on the back before filling this page) Order B7__ 5. Description of the invention (") ~ '&Quot; " Accompanied by D 3. This situation is shown under the title" Second-Order Daytime "in Figure 11. In step 1010, the new group A will share the fuses that fall on position X. Since the X price is now 1, the defective component D1 and D2 share the fuse F2. Since 01 and 1) 2¾, the two defective cabinets share a common bit 倌 "〇I on the bit position χ = ,, so this combination procedure is correct. In step 1012, the bit count value X will be Updated to 3. In Figure 11, under the heading "Second order # 代!" Is shown the result at the end of applying the second order generation to the group including defective elements D1 and D2. It will be according to steps 1014, 1016, And 1018, and for these two new groups

(分別包括瑕疵元件01/D2及D3)進行遞歸式處理。不過 .由於這櫥群在三階疊代程序開始時都没有通過步驟1004 的測試(亦即至少具有三値瑕疵元件),故對這兩値新的 群的處理程序會終丨b於步驟i 〇 2 0 (亦即不會有三階叠代 程序U 經濟部中央標準局員工消費合作社印製 (請先閱讀背面之注意事項再填寫本頁) 於二階疊代中,類似地將二階蠱代的第二群(包括三 値瑕疵元件D4, D5,及D6)當作分開的執行線而送到二 階#代中進行處理。其Μ值為三而不是小於三(步驟1〇〇4) Λ因此,此方法繼續推進到步驟1006而由這値第二群裝 配出新的Α群和新的β群(分別具有瑕疵元件D 4 , D 5 , 及OfiU如先前所述,新的A群會有INT(M/2)値瑕疵元 件。由於包栝有D4/D5/D6等瑕疵元件群中的瑕疵元件 D 4和D 6於目前的位元位置計數值X (目前的X = 1)上共用 相同的位元倌「1 !,故將瑕疵元件D4和D6組合在一起 -22- 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) A7 B7 經濟部中央標準局員工消費合作社印製 五、發明説明(- Μ )- 1 1 1 而 形 成 新 的 A 群 〇 因 此 * 新的 A群會 有 兩 個 瑕 疵 元 件 D4 . 1 1 I 和 DR 而 新 的 B 群 有 一 個備 用元件 D5 〇 這 種 情 形 是 顯 1 | 示 於 第 1 1 圖 中 標 題 厂 二 階 疊代 」底下 〇 請 先 1 I 閱 I 於 步 驟 10 10 中 新 的 A 群會 共用落 在 位 置 X 上 的 熔 絲 讀 1 背 1 Λ 由 於 現 在 X 倌 為 1 , 故瑕疵元件D 4和D 6會共用熔絲F 2 面 之 1 I η 由 於 1)1 和 D2 這 兩 値 瑕 疵 元件 都在位 元 位 置 X = =1 上 共 用 意 事 1 共 同 的 位 元 倌 厂 1 > 故這種組合程序是正確的。 於步 項 再 1 填 驟 10 1 2 中 ♦ 將 位 元 計 數 值X 更新為 3。 第1 1圖中在標 寫 本 裝 頁 1 顆 厂 二 階 疊 代 j 底 下 顯 示 的是 對包括 瑕 疵 元 件 D4和 D6的 ^' 1 群 施 行 階 # 代 程 序 結 束 時的 結果。 1 I 將 稂 據 步 驟 10 1 4 > 1016 ,及 1018 , 而 分 別 處 理 m 兩 個 1 I 新 的 群 (包括瑕疵元件D4/D5及D6)。 不 過 9 由 於 這 兩 個 1 訂 群 在 三 階 譽 代 稈 序 開 始 時 都沒 有通過 步 驟 1 0 0 4 的 測 試 ( 1 I 亦 m 至 少 具 有 三 値 瑕 疵 元 件), 故對這兩個新的群的處 1 I 理 稈 序 f-T 終 ih 於 步 驟 1020 (亦即不會有三 ί階疊代程序) 〇 1 第 12 圖 展 示 的 是 將 第 10 圔的 方法應 用 在 群 三 傾 其 位 /Λ ¥ 址 分 別 是 000 1 9 01 0 1 * 及 1111 的瑕疵 元 件 上 之 情 況 〇 為 1 了 便 於 分 配 工 作 t 加 入 設 有如 圖所示 之 號 旗 標 的 三 値 1 1 偽 造 瑕 疵 元 件 (其位址為0 0 1 0 , 0011, 及0 110) 〇 此 實 例 1 I 中 * 瑕 疵 元 件 的 原 始 群 並 未經 過分類 篩 選 〇 將 第 10 圖 步 1 1 驟 應 用 於 \-au- m 瑕 疵 元 件 的 結果 也是顯 示 於 第 12 圖 〇 1 1 第 1 3 圖 展 示 的 是 將 第 10 圖的 方法應 用 在 一 群 個 其 位 1 I 址 分 別 是 0000 * 000 1 00 10 , 0011, 01 00 9 及 1101 的 瑕 1 1 疵 元 件 上 之 情 況 〇 由 於 其 中已 經有六 m 瑕 疵 元 件 9 故 不 1 I -23- 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210·〆297公釐) A7 B7 經濟部中央標準局員工消費合作社印製 五、發明説明(>7) 1 1 需 要 加 入 任 何 偽 造 的 瑕 疵 元 件 而 使 瑕 疵 元 件 的 總 數 等 於 1 1 I 所 槔 供 備 用 元 件 的 總 數 〇 將 第 10 圖 步 驟 應 用 於 這 瑕 疵 1 件 的 結 果 ft 是 顯 示 於 第 13 圖 〇 請 先 1 1 根 據 本 發 明 的 _. β 概 念 9 在 此 提 供 了 種 高 度 精 妙 的 閱 讀 1 背 1 抟 術 以 便 在 節 省 熔 絲 巨 的 下 組 合 了 瑕 疵 元 件 之 後 將 瑕 疵 之 1 注 I 元 件 分 配 到 共 用 熔 絲 之 備 用 元 件 上 〇 於 値 實 施 例 中 9 意 牽 1 在 τπ 成 對 毎 一 群 瑕 疵 元 件 的 處 之 後 (例如在發現群的大 項 再 小 是 小 於 三 R 在 第 10 圖 中 步 驟 1 0 0 4 的 測 試 失 敗 之 後 ), 填 寫 本 '' -Ί 裝 由 群 的 第 一 値 瑕 疵 元 件 開 始 的 每 一 個 瑕 疵 元 件 只 是 分 配 頁 '〆 1 1 到 侬 序 m 增 的 標 籙 數 巨 上 〇 依 一 種 不 明 顯 的 方 式 » 所 分 1 I m 的 標 箍 數 巨 -Os- 辨 識 出 備 用 元 件 而 由 此 辨 識 出 一 個 特 殊 1 1 的 瑕 疵 元 件 r> 1 訂 為 了 展 示 、- m 値 概 念 i 故 考 慮 第 1 1 圖 的 實 例 〇 在 一 階 疊 1 I 代 稈 序 中 處 理 由 瑕 疵 元 件 D 1 , D2 j 及 D3所 構 成 的 群 之 後 1 1 • 産 生 兩 m 群 由 瑕 疵 元 件 D1和 D2 構 成 的 第 一 群 以 及 由 1 1 瑕 疵 元 件 D3 構 成 的 第 二 群 〇 在 二 階 疊 代 之 後 第 一 群 的 熔 絲 F2 是 視 為 由 (Dl/ D2) it 少·、 用 的 〇 進 入 二 階 疊 代 時 9 由 泳 1 於 瑕 疵 元 件 D 1 和 D2 構 成 的 群 大 小 是 小 於 二 * 故 、-A- m 個 群 不 1 | 能 通 過 步 驟 1 0 0 4 的 測 試 〇 根 據 所 發 明 的 分 配 技 術 9 將 1 I 瑕 疵 件 D1 分 配 到 f-TT* 標 籤 T1 而 將 瑕 疵 元 件 D2分 配 到 下 一 個 1 1 依 序 遞 增 的 標 籤 數 巨 亦 即 T2上 〇 1 | 由 於 t±r» 兀 成 了 對 由 瑕 疵 元 件 D 1和 D2 構 成 之 群 的 處 理 t 此 1 1 方 法 接 下 來 Frf 處 理 由 瑕 疵 元 件 D3所 構 成 的 群 〇 不 過 » 由 1 1 於 请 橱 群 大 小 是 小 於 三 y 故 這 個 群 也 不 能 通 過 步 驟 1004 1 I -24- 1 1 1 1 本紙張尺度適用中國國家標隼(CMS > A4規格(210X297公釐) A7 B7 經濟部中央標準局員工消費合作社印製 五、發明説明( >4 ) 1 1 的 測 試 Λ 稂 據 所 發 明 的 分 配 技 術 9 將瑕疵 元 件 D3分 配 1 1 I 到 下 一 個 依 序 遞 增 的 標 籤 數 百 上 亦 即 Τ3上。 1 1 由 於 完 成 了 對 由 瑕 疵 元 件 D 1 > D 1 > 及D3所 構 成 之 群 的 請 先 1 1 處 理 此 方 法 接 下 來 處 理 由 瑕 疵 元 件D4, D5 及 D6所 閱 讀 1 背 1 構 成 的 群 Ο 在 一 階 疊 代 程 序 中 處 理 由 瑕疵元 件 D4 9 D5 » 面 之 1 注 | D6 所 構 成 的 群 之 後 産 生 兩 値 群 由瑕疵 元 件 D4和 D6 意 1 構 成 的 第 一 群 以 及 由 瑕 疵 元 件 D5 構 成 的第二 群 〇 在 二 階 項 再 代 之 後 第 一 群 的 熔 絲 F2 是 視 為 由 (D4/ D6 ) 共 用 的 〇 供 寫 本 裝 推 λ 三 階 晷 代 時 % 由 於 瑕 疵 元 件 D4和 D6構成 的 群 大 小 是 頁 '««W〆 1 1 小 於 ~~, 故 11 個 群 不 能 通 過 步 驟 1 0 0 4的測試 〇 根 據 所 發 1 1 明 的 分 配 技 術 » F3 將 瑕 疵 元 件 D4分 配 到下一 個 依 序 遞 增 1 1 的 標 籤 數 (亦即Τ 4 ) 上 而 將 瑕 疵 元 件 D 6分配 到 下 一 個 依 1 訂 序 m 增 的 m 籤 數 g 亦 即 T5上 〇 1 I 由 於 <sSr» τη 成 了 對 由 瑕 疵 元 件 D4和 D6所 構成之 群 的 處 理 1 1 ft 方 法 接 下 來 處 理 由 瑕 疵 元 件 D5所 構成的 群 〇 不 過 9 1 1 由 於 痦 個 群 大 小 是 小 於 二 故 、-a— m 値 群 也 不能通 過 步 驟 1004 的 測 試 〇 根 據 所 發 明 的 分 配 技 術 9 將瑕疵 元 件 D5分 配 冰 1 到 下 一 個 依 序 遞 增 的 標 籤 數 巨 亦 即 Τ6上。 1 I 在 兀 成 了 對 所 有 群 的 處 理 之 後 ) 再 利用所 分 配 到 的 標 1 I 籤 數 將 每 一 値 瑕 疵 元 件 與 其 巨 m (<u«、 的 備用元 件 匹 配 在 一 1 1 起 Γ> 參 照 第 5 圖 * 分 配 到 T1 上 的 瑕 疵 元件D1 與 備 用 元 1 | 件 R 1 pc 配 在 一 起 (備用元件R 1 - R6 己 如 第5圖 所 示 作 了 編 1 | 號 ), 分配到T2上的偽造瑕疵元件D2會與備用元件R2匹配 1 1 存 一 起 r, 不 過 由 於 瑕 疵 元 件 D2只 是 一 個偽造 的 瑕 疵 元 件, 1 I -25 - 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) A7 B7 經濟部中央標準局員工消費合作社印製 五、發明説明( >4· ) - 1 1 故 不 需 要 作 任 何 的 分 配 工 作 〇 可 替 代 的 是 9 若 作 了 分配 1 1 I 刖 木 設 定 與 備 用 元 件 R2相 關 的 啓 動 熔 絲 因 而 不 用 到備 1 1 用 件 R 2 •1 請 £ 1 1 分 配 到 T3的 瑕 疵 元 件 D3 F3 與 備 用 元 件 R3 匹 配 在 起。 閱 讀 1 背 1 分 W, 到 T4的 瑕 疵 元 件 D4 1=3 與 備 用 元 件 R4匹 配 在 一 起 。分 之 1 注 | 配 到 T5的 瑕 疵 元 件 D6 與 備 用 元 件 R5 匹 配 在 起 〇 分配 意 素 1 到 Tfi的 瑕 疵 元 件 D5 與 備 用 元 件 R6匹 配 在 一 起 〇 項 再 , 第 1 2 圖 實 例 中 從 瑕 疵 元 件 到 瑕 疵 熔 絲 的 分 配 工 作 是顯 填 寫 本 裝 示 於 第 1 2 圖 内 (在標題為 「二階疊代」 此行底下是將標 頁 '〆 1 1 m 顯 示 於 瑕 疵 元 件 的 右 側 上 )〇 匹 配 的 結 果 是 D 1 / R 1 1 1 * D4 / R2 ♦ D2/ R 3 , D3 / R 4 , D6/ R 5 i 及 D5 / R6 (其中 1 1 備 用 7Π 件 R 2 * R 5 > 及 R 6不 是 未 作 分 配 就 是 其 啓 動 熔 絲未 1 訂 作 設 定 1 I 第 1 3 圖 奮 例 中 從 瑕 疵 元 件 到 瑕 疵 熔 絲 的 分 配 工 作 是顯 1 1 示 於 第 ! 3 圖 内 (在標題為 「二階覉代」 此行底下是將標 1 I 籤 顯 示 於 瑕 疵 元 件 的 右 側 上 )〇 匹 配 的 結 果 是 D 1 / R 1 • D2 / R 2 « D3/ R 3 * D5/ R 4 9 D6/ R5 t 及 D4/ R6 〇 由於 涨 1 發 現 有 個 瑕 疵 元 件 故 使 用 到 所 有 的 備 用 元 件 (亦即 ί I 有 仟 何 偽 造 的 瑕 疵 元 件 )〇 1 I 可 由 刖 述 内 容 而 得 的 優 點 是 < 本 發 明 所 提 供 的 技 術可 1 1 以 在 節 省 熔 絲 的 S 的 下 將 具 有 其 數 S 是 大 於 二 的 任 意備 1 I 用 元 件 m 合 成 任 -r*^- 備 用 之 元 件 群 0 吾 人 觀 測 到 的 是 ,可 1 以 根 據 第 8 圖 及 第 1 0 圖 中 的 步 驟 實 現 電 腦 -施行程序,而 1 1 將 發 明 件 的 節 省 技 術 在 製 造 -時間的應用以及在其替換 1 I -26 - 1 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) 經濟部中央標準局員工消費合作社印製 A7 B7 五、發明説明) 時間内的應用實現成以電腦來製成之程序,其偽依據第 8至第10圖所示之步驟^依發明所掲示内容的觀點,以 下將方法的奮現掲示成電腦程式指令是完全落在熟悉習 用稈式設計的人技術範圍内。 吾人發現對一群給定的備用元件是可以共用相當大數 目的熔絲.请導致減少了對位址熔絲數目的需求並減少 了旆行DRAM雷路時所要求的伴隨面積。將伴隨減少的位 址插梢(Utch)數目(由於每個共用熔絲只需要一値共用 的位址捅梢)及所節省下之連接到插梢及/或熔絲上的 來冏接線列人考慮時,則所達成伴隨面積竹節省更為顯 著。節省下的面精可以用於例如增加D R A Μ電路的容量。 下表展示的是可以在由備用元件構成的某些解釋用實 例中所窨現的節省情形。於第1表中,是在展示目的下 假定主陣列具有2 1D値元件而需要十個位址位元以便唯 一地定出ΐ陣列元件的位址(亦即無論是否共用,每個 備用元件都需要十個熔絲)。所節省的百分比是以每個 備用元件上含有十個位址熔絲、以及每値備用元件上含 有+個位址熔絲和一個啓動熔絲(每個備用元件上總共 含有+ —値榕絲)為基礎下計算出的。所期望的是所節 省的百分比必需有利地隨箸D R A Μ容量的增加且含有更多 備用元件、以及每個備用元件上含有更多位址熔絲的情 況而提高。 本紙張尺度適用中國國家標準(CNS ) A4規格(2丨OX297公釐) 一#= --------iri------,玎-----AW------------ (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 A7 __B7 五、發明説明(4 ) 表1 1 備用元件的數目 M = 4 M = 8 M = 16 M = 32 M = 64 M = 128 節省下的熔絲 數冃 1 5 17 49 129 321 位址熔絲的總數 (對毎個備用元 伴有10個位址熔 絲的狀況而言) 40 80 160 320 640 1280 所節省的百分比 2.5 6.25 10.625 15.3 20.15 25.1 位址熔絲的總數 (對毎®備用元 件有1.0値位址熔 苻1値啓動熔絲 的狀況而言) 44 88 176 352 704 1408 所節省的百分比 2 5.7 9.65 13.9 18,32 22.8 _ 2 8 - --------^-1裝------訂-----1¾ I .、、 (請先閱讀背面之注意事項再填寫本頁) 本紙張尺度適用中國國家標準(CNS ) A4規格(210X 297公釐) 經濟部中央標準局員工消費合作社印裝 A7 B7 五、發明説明(>7 ) 本發明是參照幾値宵施例而說明的,而有替代性、可 交換件、以及等效性的設計是落在本發明的架構下。雖 刖本發明透過實例所掲示的内容主要是參照熔絲而作的 ,然熔絲一詞也可以是指逆(a n t i )熔絲。同時,所説明 的窨旃例包含位址及啓動插梢的使用,所發明之應用在 榕絲上的熔絲節省技術也可以用在未提供這類插梢的 DRAM中。所以吾人打算將下列所附申請專利範圍解釋成 包含所有這類有替代性、可交換性、以及等效性的設計 都是落在本發明之精神及架構之内。 參者符號說明 2......主陣列元件 100. ...DRAM晶 K 1 0 2 ....主陣列 1 0 4 ....備用列或元件 2 0 2 ....解碼邏輯 2 1. 0 ....備用電路 A1 , A0.....位址位元 D1 -D6.....瑕疵元件 EF,E1-E6....啓動熔絲 E1........啓動捅楷 F 0 - F 3 ....位址熔絲 M.N......備用元件的數目 P . Q......瑕疵元件的數目 R 1 - R 6 ....備用元件 -29 - 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) τ"-裝 ; 訂 ^ (請先閲讀背面之注意事項再填寫本頁)(Including defective components 01 / D2 and D3 respectively) for recursive processing. However, since this cabinet group did not pass the test of step 1004 at the beginning of the third-order iterative process (that is, it has at least three defective elements), the processing procedure for these two new groups will end at step i. 〇 2 0 (that is, there will be no third-order iterative process U printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs (please read the precautions on the back before filling this page) In the second-order iteration, similarly the second-order iteration The second group (including the three defective elements D4, D5, and D6) is sent to the second-order #generation as a separate execution line for processing. Its M value is three instead of less than three (step 1004) Λ therefore This method continues to step 1006 and a new A group and a new β group are assembled from this second group (with defective elements D 4, D 5, and OfiU, as described previously. The new A group will have INT (M / 2) 値 Defective components. Due to the inclusion of defective components D4 / D5 / D6 in the defective component group, they are shared on the current bit position count value X (current X = 1) The same bit 倌 `` 1 !, so the defective components D4 and D6 are combined -22- This paper size applies to Chinese national standards (CNS) A4 specification (210X297 mm) A7 B7 Printed by the Consumers' Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs 5. Invention Description (-M)-1 1 1 to form a new A group. Therefore * the new A group will have two One defective component D4. 1 1 I and DR and the new B group has a spare component D5 〇 This situation is obvious 1 | shown in Figure 1 1 under the title factory second-order iteration ”please read 1 I read I In step 10 10, the new A group will share the fuses that fall on position X. Read 1 back 1 Λ Since X 倌 is now 1, the defective components D 4 and D 6 will share 1 of the fuse F 2 surface I η because 1) The two defective components 1 and D2 share the same thing at bit position X = = 1. The common bit factory 1 > Therefore, this combination process is correct. Step 1 and step 10 1 2 Medium ♦ Update the bit count value X to 3. Figure 1 1 The result of the second-order iteration j of a plant is displayed at the end of the execution of the order # program on the ^ '1 group including defective elements D4 and D6. 1 I will follow steps 10 1 4 > 1016, and 1018, while m two new 1 I groups (including defective components D4 / D5 and D6) will be processed separately. However, 9 Because these two 1-order groups failed the test of step 1 0 0 at the beginning of the third-order reputation order (1 I also has at least three-dimensional defective elements), the At 110, the sequence fT ends ih in step 1020 (that is, there will be no three-th order iterative procedures) 〇1 Figure 12 shows the application of the 10th method to the group of three inclination / Λ ¥ address respectively Yes 000 1 9 01 0 1 * and 1111 on the defective component. 0 is 1 to facilitate the assignment of work. T Adding three flags with a flag as shown in the figure 1 1 Forging the defective component (its address is 0 0 1 0 , 0011, and 0 110) 〇 In this example 1 * the original group of defective components has not been classified and filtered 〇 Applying step 1 1 of Figure 10 to \ -au- m defective components is also shown in Figure 12 〇1 1 Figure 1 3 shows the method of Figure 10 applied to a group of bits whose 1 I addresses are 0000 * 000 1 00 10, 0011, 01 00 9 and 1101 defects 1 1 Condition on defective components 〇 There are already 6 m defective components 9 so not 1 I -23- 1 1 1 This paper size applies Chinese National Standard (CNS ) A4 specification (210 · 〆297 mm) A7 B7 Printed by the Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs 5. Description of the invention (> 7) 1 1 It is necessary to add any forged defective components so that the total number of defective components equals 1 1 I The total number of spare parts available. The result of applying step 10 to this defect 1 piece ft is shown in FIG. 13 Please first 1 1 According to the _. Β concept 9 of the present invention, a height is provided here Intensive reading 1 Back 1 operation to save defective fuses after combining defective components 1 Note I Components are allocated to spare components that share fuses. Pull 1 after τπ pairs a group of defective components (for example, after finding that the large term of the group is smaller than three R, after the test of step 1 0 0 4 in Figure 10 fails), fill out this ''-Ί 装Each defective component starting from the first defective component of the group is only allocated on the page '〆1 1 to the order of the number of increments of the non-negative sequence m is large. In an inconspicuous way »the number of divided 1 I m marks is huge -Os- Identified a spare component and thus identified a special 1 1 defective component r > 1 is intended to show,-m 値 concept i, so consider the example of Figure 1 1 in the first-order stack 1 I generation sequence After processing the group consisting of defective elements D 1, D2 j and D3 1 1 • Generate two m groups consisting of the first group consisting of defective elements D1 and D2 and the second group consisting of 1 1 defective elements D3. The first group of fuses F2 after generation is regarded as (Dl / D2) It is less. When using 〇 to enter the second-order iteration 9 The size of the group formed by swimming 1 with defective elements D 1 and D2 is less than two. Therefore, -A- m groups are not 1 | Can pass step 1 0 0 4 Test 〇 Assign 1 I defective part D1 to f-TT * label T1 and assign defective component D2 to the next 1 1 according to the invented distribution technology 1 1 The number of sequentially increasing labels is huge on T2 〇1 | Since t ± r »becomes the processing of the group consisting of defective elements D 1 and D2. This 1 1 method is followed by Frf processing the group of defective elements D3. However, the size of the group by 1 1 is less than three. y Therefore, this group cannot pass step 1004 1 I -24- 1 1 1 1 This paper size applies to the Chinese national standard (CMS > A4 size (210X297 mm) A7 B7 Printed by the staff consumer cooperative of the Central Standards Bureau of the Ministry of Economic Affairs 2. Test of Invention (> 4) 1 1 According to the invention Assignment technology 9 assigns defective element D3 1 1 I to the next successively increasing number of tags, which is hundreds of TT3. 1 1 As a result of the group consisting of defective elements D 1 > D 1 > and D3, please proceed first 1 1 This method is followed by the group consisting of defective elements D4, D5 and D6 read 1 back 1 〇 After processing the group consisting of defective elements D4 9 D5 in the first-order iterative program, the group consisting of Note 1 | D6 produces two first groups consisting of defective elements D4 and D6, and the first group consisting of defective elements D5 The second group of 0. After the second-order term is replaced, the fuse F2 of the first group is considered to be shared by (D4 / D6). 0 For writing, the third-order generation of λ is a group consisting of defective components D4 and D6. The size is the page '«« W〆1 1 is less than ~~, so 11 groups cannot pass the test of step 1 0 0 4 0 according to the distribution technology issued by 1 1 »F3 assigns the defective component D4 to the next sequential increase 1 1 on the number of labels (that is, T 4) and will be defective Piece D 6 is assigned to the next m order number m which increases by 1 g, that is, T1 on T5 〇 I I < sSr »τη becomes the treatment of the group composed of defective elements D4 and D6 1 1 ft method Next, the group consisting of the defective component D5 is processed. However, 9 1 1 because the size of the group is less than two, the -a-m group cannot pass the test of step 1004. According to the invented distribution technology 9, the defective component is D5 allocates ice 1 to the next successively increasing number of labels, which is T6. After 1 I becomes the processing of all groups), the number of assigned 1 I tickets is used to match each defective component with its spare component of m (< u «,) from 1 1 Γ > Refer to Figure 5 * Defective component D1 assigned to T1 and spare element 1 | R 1 pc are matched together (spare components R 1-R6 have been numbered 1 | as shown in Figure 5) and assigned to T2 The counterfeit defective component D2 will be matched with the spare component R2 1 1 and stored together, but because the defective component D2 is only a fake defective component, 1 I -25-1 1 1 This paper size applies to China National Standard (CNS) A4 specifications (210X297 mm) A7 B7 Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs. 5. Description of the invention (> 4 ·)-1 1 Therefore, no distribution work is required. 0 can be replaced. 9 Allocation 1 1 I Tochigi sets the start-up fuse related to spare element R2, so spare 1 1 is not required. R 2 • 1 Please £ 1 1 Defective component D3 assigned to T3 F3 matches spare component R3. Read 1 back 1 minute W, Defective component D4 to T4 1 = 3 matches spare component R4. 1 in 1 Note | Defective component D6 assigned to T5 matches spare component R5. Defective component D5 assigned to elements 1 to Tfi is matched with spare component R6. Item 1, then from the defective component to the defective fuse in the example in Figure 1 2 The assignment work is to fill out this installation shown in Figure 12 (under the title "Second-Order Iteration", the tab '〆1 1 m is displayed on the right side of the defective component). The matching result is D. 1 / R 1 1 1 * D4 / R2 ♦ D2 / R 3 , D3 / R 4 , D6 / R 5 i and D5 / R6 (of which 1 1 spare 7Π pieces R 2 * R 5 > and R 6 are not inactive Allocation is its start-up fuse. 1 Custom setting 1 I Chapter 1 3 Figure Fen The assignment from defective components to defective fuses is shown in the figure 1 1 shown in Figure 3 (under the line titled "Second-Order Generation", the label 1 I is displayed on the right side of the defective component). The result is D 1 / R 1 • D2 / R 2 «D3 / R 3 * D5 / R 4 9 D6 / R5 t and D4 / R6 〇 Due to the increase of 1 found a defective component, all spare components are used (also That is, I have no forged defective components) 〇1 The advantage that can be obtained from the description is that the technology provided by the present invention can 1 1 to have a number S that is greater than S, which saves fuses. Arbitrary backups of two I I use component m to synthesize any -r * ^-spare component group 0 What we have observed is that we can implement the computer-executing program according to the steps in Figure 8 and Figure 10, and 1 1 The application of invention-saving techniques in manufacturing-time and its replacement 1 I -26-1 1 1 1 Paper size applies Chinese National Standard (CNS) A4 specification (210X297 mm) Printed by the Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Description of the invention) Application within time is realized as a computer-made program, which has a false basis The steps shown in Figs. 8 to 10 ^ From the viewpoint of the content shown in the invention, the following shows the method's emergence as a computer program instruction is completely within the technical scope of a person familiar with conventional design. I have found that for a given group of spare components, a relatively large number of fuses can be shared. Please reduce the number of address fuses and the accompanying area required when walking the DRAM circuit. Reduce the number of address pins (since each shared fuse requires only one shared address pin) and the saved wiring to the pins and / or fuses When people think about it, the bamboo savings achieved along with the area are even more significant. The saved noodles can be used, for example, to increase the capacity of a DRAM circuit. The following table shows the savings that can be realized in some interpreted examples of spare parts. In the first table, it is assumed that the main array has 2 1D 値 elements for display purposes. Ten address bits are required to uniquely determine the address of the ΐ array elements (that is, each spare element is shared regardless of whether it is shared or not). Needs ten fuses). The percentage saved is ten address fuses on each spare element, plus + address fuses and a startup fuse on each spare element (total + + on each spare element) ). It is desirable that the percentage of savings must be advantageously increased as the 箸 DR AM capacity increases and more spare components are included, and more address fuses are included on each spare component. This paper size applies to Chinese National Standard (CNS) A4 specification (2 丨 OX297 mm) ## -------- iri ------, 玎 ----- AW ----- ------- (Please read the notes on the back before filling this page) Printed by the Consumer Cooperatives of the Central Bureau of Standards of the Ministry of Economic Affairs A7 __B7 V. Description of the invention (4) Table 1 1 Number of spare parts M = 4 M = 8 M = 16 M = 32 M = 64 M = 128 The number of fuses saved 冃 1 5 17 49 129 321 The total number of address fuses (for a spare cell with 10 address fuses) Speech) 40 80 160 320 640 1280 Percentage saved 2.5 6.25 10.625 15.3 20.15 25.1 Total number of address fuses (for 毎 ® spare components with 1.0 値 address fuse 苻 1 値 fuse activated) 44 88 176 352 704 1408 Percent of savings 2 5.7 9.65 13.9 18,32 22.8 _ 2 8--------- ^-1 equipment -------- order ----- 1¾ I. ,, (please Please read the notes on the back before filling this page) This paper size is applicable to China National Standard (CNS) A4 (210X 297 mm) Printed on A7 B7 by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs 5. Description of the invention (> 7) This invention is a reference As described in the following examples, alternative, interchangeable, and equivalent designs fall within the framework of the present invention. Although the content shown in the present invention through examples is mainly made with reference to a fuse, the term fuse may also refer to an inverse (a n t i) fuse. At the same time, the illustrated examples include the use of addresses and boot pins, and the invented fuse-saving technology applied to Fibre can also be used in DRAMs that do not provide such pins. Therefore, I intend to interpret the scope of the following patent application as including all such alternative, interchangeable, and equivalent designs that fall within the spirit and framework of the present invention. Reference symbol description 2 ... Main array element 100 ... DRAM crystal K 1 0 2 .... Main array 1 0 4 .... Spare column or element 2 0 2 .... Decode Logic 2 1. 0 .... Backup circuits A1, A0 ..... Address bits D1 -D6 ..... Defective elements EF, E1-E6 .... Enable fuse E1 ... .... Startup script F 0-F 3 .... Address fuse MN ..... Number of spare parts P. Q ...... Number of defective parts R 1-R 6 .. .. Spare Parts-29-This paper size applies to China National Standard (CNS) A4 (210X297mm) τ "-Packing; Order ^ (Please read the precautions on the back before filling this page)

Claims (1)

經濟.部中央標準局員工消費合作社印製 A8 1 B8 C8 D8 ^、申請專利範圍 1. 一種於記億體陣列中製成許多共用熔絲之第一備用元 件時降低位址熔絲數目所用之方法,此方法之特徵包 栝: 確定該許多第一備用元件中備用元件的數目是否至 少有三艏;以及‘ 若該許多第一備用元件中該備用元件的數目是至少 有三値.刖依下列步驟進行最佳化: 由該許多第一備用元件形成第一備用元件群, '1 由該許多第一備用元件形成第二備用元件群,該第 二備用元件群輿該第一備用元件群之間是互斥的,以及 提供一橱第一熔絲以供該第一備用元件群中的各備 用元件共用,該第一熔絲會形成該第一備用元件群中 各備用元件的最高階位址熔絲。 2 .如申請專利範圍第1項之方法,其中該記億體陣列代 表的皋一値動態隨機存取記億體所構成的陣列。 3.如申請專利範圍第1項之方法,其中該第一備用元件 群是由INT(M/2)個備用元件所構成的,而其中Μ代 表的是該許多第一備用元件中該備用元件的數目,該 許多第一備用元件中的剩餘備用元件則屬於該第二備 用元件群。 如由請專利範圍第3項之方法,其中該最佳化程序也 包栝: 若該第二備用元件群中備用元件的數目是至少有三 - -· ·.. 脑,則再依下列步驟進行位址熔絲的最佳化: -30- 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) nn m^i m m^i Bl·—— tn^ -ft —1 tn ml ·ϋϋ m^i 一一 nn nn nn —^ϋ ^ ,/ ¾ 、v$'、系 (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印製 六、申請專利範圍 . 由該第二備用元件群形成第三備用元件群; 由該第二備用元件群形成第四備用元件群,該第四 備用元件群與該第三備用元件群之間是互斥的,以及 掃供一龆第二熔絲以供該第三備用元件群中的各備 用元件共用,該第二熔絲會形成該第一備用元件群中 各備用元件的僅次於最高階之位址熔絲。 5.如由請專利範圍第4項之方法,其中該第三備用元件 群是由TNT(N/2)個備用元件所構成的,而其中N代表 的是該第二備用元件群中該備用元件的數目,該第二 備用元件群中的剩餘備用元件則屬於該第四備用元件 TtW. WP A 6 . —種用於具有處理器的電腦中的電腦-施行方法,以減 少於記億體陣列中製成許多共用熔絲之第一備用元件 時所用位址熔絲數目,此種電腦-施行方法之特徵包括: 確定該許多第一備用元件中備用元件的數目是否至 少有三値;以及 若該許多第一備用元件中該備用元件的數目是至少 有三Μ ,則依下列步驟進行位址熔絲的最佳化: 利用該處理器由該許多第一備用元件形成第一備用 元件群, 利用該處理器由該許多第一備用元件形成第二備用 元件群,該第二備用元件群與該第一備用元件群之間 是互斥的,以及 利用該處理器設計一個第一熔絲以供該第一備用元 -3 1- 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) --------i------IT-----W線 (請先閲讀背面之注意事項再填寫本頁) 經濟部中央標率局員工消費合作社印製 A8 B8 C8 D8 六、申請專利範圍 件群中的各備用元件共用,該第一熔絲會形成該第一 備用元件群中各備用元件的最高階位址熔絲。 7 .如由請專利範圍第6項之電腦-施行方法,其中該第 一備用元件群是由INT(M/2)個備用元件所構成的, 而其中Μ代表的是該許多第一備用元件中該備用元件 的數目,該許多第一備用元件中的剩餘備用元件則鼷 於該第二備用元件群。 8 ·如申請專利範圍第7項之電腦-施行方法,其中再依 下列步驟淮行該位址熔絲的最佳化: 若該第二備用元件群中備用元件的數目是至少有三 個,刖再依下列步驟進行位址熔絲的最佳化: 利用該處理器由該第二備用元件群形成第三備用元 Λ 件群, 利用該處理器由該第二備用元件群形成第四備用元 件群,該第四備用元件群與該第三備用元件群之間是 π斥的,以及 利用該處理器設計一値第二熔絲以供該第三備用元 件倉中的各備用元件共用,該第二熔絲會形成該第一 備用元件群中各備用元件的僅次於最高階之位址熔絲。 9 .如申請專利範圍第8項之電腦-施行方法,其中該第 三備用元件群是由ΙΝΤ(Ν/2)個備用元件所構成的, 而其中Ν代表的是該第二備用元件群中該備用元件的 數目.該第二備用元件群中的剩餘備用元件則屬於該 第四備用元件群。 本紙張尺度適用中國國家標準(CNS ) Α4規格(210X297公釐) ---------'Ί裝------訂-----線 (請先閲讀背面之注意事項再填寫本頁) A8 Βδ C8 D8 經濟部中央標準局員工消費合作社印製 申請專利範圍 10. —種記億體陣列,具有許多共用熔絲之第一備用元 件取取代記億體陣列中的瑕疵元件,其待徵包括: 一個第一懷絲; 一痼由該許多第一備用元件所形成之第一備用元件 群,該第一備用元件群將共用的該第一熔絲當作它們 的最高階位址熔絲;以及 由該許多第一備用元件形成第二備用元件群,該第 二備用元件群與該第一備用元件群之間是互斥的。 η .如由請專利範圍第1 〇項之記憶體陣列,其中該記億 體陣列代表的是一痼動態隨機存取記億體所構成的陣 列,、 1 2.如申請專利範圍第1 〇項之記億體陣列,其中該第一 備用元件群是由ΙΝΤ(Μ/2)個備用元件所構成的,而 其中Μ代表的是該許多第一備用元件中該備用元件的 數目,該許多第一備用元件中的剩餘備用元件則屬於 該第二備用元件群。 13.如申請專利範圍第12項之記億體陣列,其中又包括 一値第二熔絲,且其中由備用元件所構成的該第二備 用元件群包括: 一個第三備用元件群,該第三備用元件群將共用的 該第二熔絲當作它們的次高階位址熔絲;以及 一値第四備用元件群,該第四備用元件群與該第三 備用元件群之間是互斥的。 1 4 .如申請專利範圍第1 3項之記億體陣列,其中該第三 -33- 本紙張尺度適用中國國家標準(CNS ) Α4见格(210Χ297公釐) --------'Ί 裝------訂-----<一:〕線 (請先閱讀背面之注意事項再填寫本頁) 經濟部中央標準局員工消費合作社印裝 A8 B8 C8 D8 ^、申請專利範圍 備用元件群是由IHT(N/2)個備用元件所構成的,而 其中K代表的是該第二備用元件群中該備用元件的數 目,該第二備用元件群中的剩餘備用元件則屬於該第 四備用元件群。 15.—種用於電腦中的電腦-施行方法,其中傺將記億體 陣列中的許多第一瑕疵元件取代成許多共用熔絲之第 一備用元件,此方法之特徽包括: 確定該許多第一瑕疵元件中瑕疵元件的數目是否小 於該許多共用熔絲之第一備用元件中備用元件的數目; 若該許多第一瑕疵元件中該瑕疵元件的數目是小於 該許多共用熔絲之第一備用元件中該備用元件的數目 ,則於該許多第一瑕疵元件中加入許多偽造的瑕疵元 件以便將該許多第一瑕疵元件中瑕疵元件的數目增加 到該第一共用熔絲之第一備用元件中該備用元件的敷 目; 確定該許多第一瑕疵元件中該瑕疵元件的數目是否 至少有三値;以及 若該許多第一瑕疵元件中該瑕疵元件的數目是至少 有三個,則利用執行下列步驟而將該許多第一瑕疵元 件中的個別元件取代成該許多第一備用元件中的値別 元件: 由該許多第一瑕疵元件形成第一瑕疵元件群,該第 一瑕疵元件群中的各瑕疵元件是於最高階位址位元置 上共用相同的位元值, 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) m m 1^1 m n^i nn ΙΛ .n ml ml —^ϋ am in I— n I-I n nn^ ^ 「-¾ i VI、4' (請先閱讀背面之注意事項再填寫本頁) ABCD 經濟部中央標準局員工消費合作社印製 々、申請專利範圍 由該許多第一備用元件形成第二備用元件,該第二 備用元件群與該第一備用元件群之間是互斥的,以及 將該第一瑕疵元件群中的個別元件分配到該第一備 用元件群中的個別元件以便進行取代。 1 6 .如由請專利範圍第1 5項之電腦-施行方法,其中該第 一備用元件群是由INT (P/ 2)個備用元件所構成的,而 其中P代表的是該許多第一備用元件中該備用元件的 數目,該許多第一備用元件中的剩餘備用元件則屬於 該第二備用元件群。 1 7 .如申請專利範圍第1 6項之電腦-施行方法,其中該取 代程序又包活: 若該第二瑕疵元件群中瑕疵元件的數目是至少有三 橱,則再進行下列步驟: 由該第二瑕疵元件群形成第三瑕疵元件群,該第三 瑕疵元件群中的瑕疵元件是於次高階位址位元位置上 共用一個相同的位元值, u 由該第二瑕疵元件群形成第四瑕疵元件群,該第四 瑕疵元件群與該第三瑕疵元件群之間是互斥的,以及 將該第三瑕疵元件群中的個別元件分配到該第一備 用元件群中的値別元件以便進行取代。 1 8 .如申請專利範圍第1 7項之電腦-施行方法,其中該第 三備用元件群是由INT(Q/2)個備用元件所構成的, 而其中Q代表的是該第二備用元件群中該備用元件的 數目.該第二備用元件群中的剩餘備用元件則屬於該 -35 - 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) --------裝丨| (請先閲讀背面之注意事項再填寫本頁) 訂 經濟部中央標準局員工消費合作社印製 A8 B8 C8 D8 々、申請專利範圍 第四備用元件群。 1 9 .如由請專利範圍第1 5項之電腦-施行方法,其中該記 億體陣列代表的是一痼動態隨機存取記億體所構成的 陳列,、 2 0 . —種使記億體陣列中的許多第一瑕疵元件取代成許 多共用熔絲之第一備用元件所用之方法,此方法之待 徽包括: 確定該許多第一瑕疵元件中瑕疵元件的數目是否小 於該許多共用熔絲之第一備用元件中備用元件的數目; 若該許多第一瑕疵元件中該瑕疵元件的數目是小於 該許多共用熔絲之第一備用元件中該備用元件的數目 .刖於該許多第一瑕疵元件中加入許多偽造的瑕疵元 件以便將該許多第一瑕疵元件中瑕疵元件的數目增加 到該許多共用熔絲之第一備用元件中該備甩元件的數 S ; 確定該記億體陣列中該許多第一瑕疵元件中該瑕疵 元件的數目是否至少有三個;以及 若該許多第一瑕疵元件中該瑕疵元件的數目是至少 有三痼,刖利用執行下列步驟而將該許多第一瑕疵元 件中的個別元件取代成該許多第一備用元件中的個別 元件: 由該許多第一瑕疵元件形成笫一瑕疵元件群,該第 一瑕疵元件群中的各接線元件是於最高階位址位元位 詈L·共用相同的位元值, -36 - 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) --------^ —裝------訂-----{ 森 ..V (請先閲讀背面之注意事項再填寫本頁) A8 Βδ C8 D8 六、申請專利範園 由該許多第一備用元件形成第二備用元件群,該第 二備用元件群與該第一備甩元件群之間是互斥的,以及 將該第一瑕疵元件群中的個別元件分配到該第一備 用元件群中的値別元件以便進行取代。 21.如申請專利範園第20項之方法,其中該第一備用元件 群是由T N T ( P / 2 )個備用元件所構成的,而其中P代表 的是該許多第一備用元件中該備用元件的數目,該許 多第一備用元件中的剩餘備用元件則屬於該第二備用 元件群。 2 2 .如申請專利範圍第2 1項之方法,其中該取代程序又 包栝: 若該第二瑕疵元件群中瑕疵元件的數目是至少有三 櫥.刖再谁行下列步驟: 由該第二瑕疵元件群形成第三瑕疵元件群,該第三 瑕疵元件群中的瑕疵元件是於次高階位址位元位置上 共用一橱相同的位元值, 由該第二瑕疵元件群形成第四瑕疵元件群,該第四 薛疵元件群與該第三瑕疵元件群之間是互斥的,以及 經濟部中央標準局員工消費合作社印製 (請先聞讀背面之注意事項再填寫本頁) 將該第三瑕疵元件群中的個別元件分配到該第一備 用元件群中的値別元件以便進行取代。 2 3.如申請專利範圍第22項之方法,其中該第三備用元件 群是由INT(Q/2)個備用元件所構成的,而其中Q代 表的是該第三備用元件群中該備用元件的數目,該第 -37- 本紙張尺度逋用中國國家標準(CNS ) A4規格(210X297公釐) A8 B8 C8 D8 々、申請專利範圍 二備用元件群中的剩餘備1元:件刖屬於該第四備用元 件群。 2 4 .如由請專利範圍第2 0項之方法,其中該記億體陣列 代寿的是一個共用熔絲之記億體所構成的陣列。 (請先閱讀背面之注意事項再填寫本頁) .裝_ 、1T 經濟部中央標準局員工消費合作社印製 -38- 本紙張尺度適用中國國家標準(CNS ) A4%格(210X297公釐)Economy. Printed by A8 1 B8 C8 D8 ^, Consumer Cooperatives of the Central Standards Bureau of the Ministry of Commerce, patent application scope 1. A method for reducing the number of address fuses when the first spare component of many common fuses is made in a memory array Method, the method includes the following features: determining whether the number of spare elements in the plurality of first spare elements is at least three; and 'if the number of spare elements in the plurality of first spare elements is at least three, follow the steps below Optimizing: forming a first spare element group from the plurality of first spare elements, '1 forming a second spare element group from the plurality of first spare elements, and the second spare element group between the first spare element group Are mutually exclusive, and a cabinet of first fuses is provided for sharing among the spare components in the first spare component group, and the first fuse forms the highest order address of each spare component in the first spare component group Fuse. 2. The method according to item 1 of the scope of patent application, wherein the array composed of the random random access memory of the memory array represented by the array of the memory array. 3. The method according to item 1 of the patent application scope, wherein the first spare component group is composed of INT (M / 2) spare components, and M represents the spare component among the many first spare components. , The remaining spare elements in the many first spare elements belong to the second spare element group. For example, if the method of claim 3 is adopted, the optimization procedure also includes: If the number of spare components in the second spare component group is at least three--· · .. brain, then follow the steps below Optimization of address fuse: -30- This paper size applies to Chinese National Standard (CNS) A4 (210X297 mm) nn m ^ imm ^ i Bl · —— tn ^ -ft —1 tn ml · ϋϋ m ^ i 一一 nn nn nn — ^ ϋ ^, / ¾, v $ ', Department (Please read the notes on the back before filling out this page) Printed by the Consumers' Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs 6. Scope of patent application. The second spare component group forms a third spare component group; the second spare component group forms a fourth spare component group, the fourth spare component group and the third spare component group are mutually exclusive, and the supply is scanned. A stack of second fuses is shared by all the spare components in the third spare component group, and the second fuse will form an address fuse of the spare components in the first spare component group only after the highest order. 5. If the method of claim 4 is adopted, the third spare component group is composed of TNT (N / 2) spare components, and N represents the spare component in the second spare component group. The number of components, and the remaining spare components in the second spare component group belong to the fourth spare component TtW. WP A 6. —A computer for a computer with a processor—implementation method to reduce the number of billions The number of address fuses used to make the first spare component of the common fuse in the array. The computer-implementation method includes: determining whether the number of spare components in the many first spare components is at least three; and if The number of the spare elements in the plurality of first spare elements is at least three M, then the address fuse is optimized according to the following steps: using the processor to form a first spare element group from the plurality of first spare elements, using The processor forms a second spare element group from the plurality of first spare elements, the second spare element group and the first spare element group are mutually exclusive, and a first A fuse for the first spare element-3 1- This paper size applies to China National Standard (CNS) A4 specification (210X297 mm) -------- i ------ IT --- --W line (please read the precautions on the back before filling this page) Printed by A8 B8 C8 D8 of the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economy The filaments will form the highest order address fuses of each spare element in the first spare element group. 7. If the computer-implementation method of item 6 of the patent scope is adopted, wherein the first spare component group is composed of INT (M / 2) spare components, and M represents the many first spare components The number of the spare elements in the plurality of first spare elements falls in the second spare element group. 8 · If the computer-implementation method of item 7 of the scope of patent application, wherein the optimization of the address fuse is performed according to the following steps: If the number of spare components in the second spare component group is at least three, 刖Then, the address fuse is optimized according to the following steps: using the processor to form a third spare element group from the second spare element group, and using the processor to form a fourth spare element from the second spare element group; Group, the fourth spare component group and the third spare component group are π-excluded, and a second fuse is designed by the processor to be shared by all spare components in the third spare component bin, the The second fuse forms an address fuse of the spare elements in the first spare element group after the highest order. 9. The computer-implementing method according to item 8 of the scope of patent application, wherein the third spare component group is composed of INT (N / 2) spare components, and N represents the second spare component group. The number of the spare components. The remaining spare components in the second spare component group belong to the fourth spare component group. This paper size applies to China National Standard (CNS) Α4 specification (210X297 mm) --------- 'outfitting ------ order ----- line (Please read the precautions on the back first (Fill in this page again.) A8 Βδ C8 D8 Printed by the Consumers' Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs. The scope of patent application is 10.-A kind of memory array, the first spare with many common fuses to replace the defects in the memory array Components, whose characteristics include: a first white wire; a first spare component group formed by the plurality of first spare components, the first spare component group taking the first fuse in common as their highest Order address fuses; and a second spare element group formed by the plurality of first spare elements, the second spare element group and the first spare element group being mutually exclusive. η. For example, the memory array of item 10 of the patent scope, where the memory array represents a memory array composed of a dynamic random access memory, 1 2. As the scope of patent application No. 10 The billion-body array of the item, wherein the first spare element group is composed of INT (M / 2) spare elements, and M represents the number of the spare elements among the many first spare elements, and the many The remaining spare elements in the first spare element belong to the second spare element group. 13. The billion-body array of item 12 of the patent application scope, which further includes a stack of second fuses, and the second spare element group composed of spare elements includes: a third spare element group, the first The three spare component groups use the shared second fuse as their next higher-order address fuse; and a fourth spare component group, the fourth spare component group and the third spare component group are mutually exclusive. of. 1 4. If the billion-body array of the 13th item in the scope of patent application, the third -33- this paper size applies the Chinese National Standard (CNS) Α4 see the grid (210 × 297 mm) -------- 'Ί 装 ------ Order ----- < one:] line (please read the precautions on the back before filling this page) Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs A8 B8 C8 D8 ^, Patent application scope The spare component group is composed of IHT (N / 2) spare components, where K represents the number of the spare components in the second spare component group, and the remaining spares in the second spare component group The components belong to the fourth spare component group. 15. A computer-implementation method for a computer, wherein a plurality of first defective components in a memory array are replaced by a plurality of first spare components of a common fuse, and the features of this method include: determining the many Whether the number of defective components in the first defective component is less than the number of spare components in the first spare component of the many common fuses; if the number of defective components in the many first defective components is less than the first of the many common fuses The number of the spare components in the spare component, a plurality of forged defective components are added to the many first defective components in order to increase the number of defective components in the first defective components to the first spare component of the first common fuse Determine the number of the defective components in the first defective component, and determine whether the number of the defective components is at least three; and if the number of the defective components in the first defective components is at least three, perform the following steps: And replacing individual components in the many first defective components with individual components in the many first spare components: Multiple first defective components form a first defective component group, and each defective component in the first defective component group shares the same bit value on the highest-order address bit, and this paper standard applies Chinese National Standard (CNS) A4 size (210X297 mm) mm 1 ^ 1 mn ^ i nn ΙΛ .n ml ml — ^ ϋ am in I— n II n nn ^ ^ "-¾ i VI, 4 '(Please read the precautions on the back first (Fill in this page) Printed by the ABCD Employee Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs, the patent application scope is formed by the many first spare components to form a second spare component, and the second spare component group and the first spare component group are interactive. And the individual components in the first defective component group are assigned to the individual components in the first spare component group so as to be replaced. 1. If the computer-implementation method of item 15 of the patent scope is applied, wherein The first spare element group is composed of INT (P / 2) spare elements, and P represents the number of the spare elements among the many first spare elements, and the remaining spares among the many first spare elements. Component belongs to the Spare component group 17. If the computer-implementation method of item 16 of the patent application scope, wherein the replacement procedure is also active: If the number of defective components in the second defective component group is at least three cabinets, then perform the following Steps: A third defective component group is formed from the second defective component group. The defective components in the third defective component group share the same bit value at the next higher-order address bit position. The component group forms a fourth defective component group, the fourth defective component group and the third defective component group are mutually exclusive, and individual components in the third defective component group are allocated to the first spare component group. Unique components for replacement. 18. The computer-implementation method according to item 17 of the scope of patent application, wherein the third spare component group is composed of INT (Q / 2) spare components, and Q represents the second spare component The number of spare components in the group. The remaining spare components in the second spare component group belong to the -35-This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -------- (Please read the precautions on the back before filling out this page) Order the A8 B8 C8 D8 printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs 々, the fourth spare component group for the scope of patent application. 19. If the computer-implementation method of item 15 of the patent scope is applied, wherein the array of billions of digits represents an array of dynamic random access digits of billions of digits, 2... A method for replacing many first defective components in a bulk array with first spare components of a plurality of shared fuses, the method of which includes: determining whether the number of defective components in the plurality of first defective components is less than the plurality of shared fuses The number of spare components in the first spare component; if the number of the defective components in the many first defective components is less than the number of the spare components in the first spare components of the many common fuses; Adding many forged defective components to the component in order to increase the number of defective components in the many first defective components to the number S of the spare components in the first spare component of the many common fuses; Whether the number of the defective elements in the many first defective elements is at least three; and if the number of the defective elements in the many first defective elements is at least three执行 replacing the individual components in the plurality of first defective components with the individual components in the plurality of first spare components by performing the following steps: forming a defective component group from the plurality of first defective components, the first defective component group Each wiring element in the highest order address bit 詈 L · shares the same bit value, -36-This paper size applies to China National Standard (CNS) A4 specification (210X297 mm) ------ -^ — 装 ------ Order ----- {Mori..V (Please read the precautions on the back before filling this page) A8 Βδ C8 D8 The spare components form a second spare component group, the second spare component group and the first spare component group are mutually exclusive, and individual components in the first defective component group are allocated to the first spare component group. In order to replace the different elements. 21. The method according to item 20 of the patent application park, wherein the first spare element group is composed of TNT (P / 2) spare elements, and P represents the spare among the many first spare elements. The number of components, and the remaining spare components in the many first spare components belong to the second spare component group. 2 2. The method according to item 21 of the scope of patent application, wherein the replacement procedure further includes: If the number of defective components in the second defective component group is at least three cabinets, who can perform the following steps: From the second The defective element group forms a third defective element group. The defective elements in the third defective element group share the same bit value at the next higher-order address bit position, and a fourth defect is formed by the second defective element group. Component group, the fourth defective component group and the third defective component group are mutually exclusive, and printed by the Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs (please read the precautions on the back before filling out this page) Individual components in the third defective component group are assigned to individual components in the first spare component group for replacement. 2 3. The method according to item 22 of the scope of patent application, wherein the third spare component group is composed of INT (Q / 2) spare components, and Q represents the spare component in the third spare component group. The number of components, this section -37- This paper size uses the Chinese National Standard (CNS) A4 specification (210X297 mm) A8 B8 C8 D8 々, the remaining spare parts in the spare component group of the scope of patent application 1 yuan: pieces 刖The fourth spare element group. 24. If the method of claiming the 20th item of the patent scope is adopted, wherein the array of billions of bodies is replaced by an array of shared billions of bodies. (Please read the precautions on the back before filling out this page.) _ _, 1T Printed by the Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs -38- This paper size applies to China National Standard (CNS) A4% grid (210X297 mm)
TW087109644A 1997-06-30 1998-06-17 Improved techniques for reducing redundant element fuses in a dynamic random access memory array TW384475B (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US08/884,854 US5831917A (en) 1997-06-30 1997-06-30 Techniques for reducing redundant element fuses in a dynamic random access memory array

Publications (1)

Publication Number Publication Date
TW384475B true TW384475B (en) 2000-03-11

Family

ID=25385566

Family Applications (1)

Application Number Title Priority Date Filing Date
TW087109644A TW384475B (en) 1997-06-30 1998-06-17 Improved techniques for reducing redundant element fuses in a dynamic random access memory array

Country Status (3)

Country Link
KR (1) KR19990007466A (en)
DE (1) DE69826075D1 (en)
TW (1) TW384475B (en)

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR950001837B1 (en) * 1992-07-13 1995-03-03 삼성전자주식회사 Row redundancy circuit sharing fuse box
KR950004623B1 (en) * 1992-12-07 1995-05-03 삼성전자주식회사 Semiconductor memory device with improved redundancy efficiency
US5646896A (en) * 1995-10-31 1997-07-08 Hyundai Electronics America Memory device with reduced number of fuses

Also Published As

Publication number Publication date
DE69826075D1 (en) 2004-10-14
KR19990007466A (en) 1999-01-25

Similar Documents

Publication Publication Date Title
TWI229197B (en) Built-in spare row and column replacement analysis system for embedded memories
US6262935B1 (en) Shift redundancy scheme for wordlines in memory circuits
CN109840432A (en) The method for generating security key for IC apparatus
KR100630519B1 (en) Dimension programmable fusebanks and methods for making the same
DE102006048856B4 (en) A method of operating an IC semiconductor memory device and an IC semiconductor memory device
DE10206720A1 (en) System and method for operating a programmable counter for column failure with redundancy assignment
TW548653B (en) Semiconductor memory device having redundancy memory circuit
WO2000011554A1 (en) A system and method for identification of transformation of memory device addresses
WO2002089147A3 (en) Circuit and method for memory test and repair
KR960019312A (en) Random Access Memory with Flexible Array Redundancy
US11200962B2 (en) Memory devices having spare column remap storages and methods of remapping column addresses in the memory devices
US6879529B2 (en) Semiconductor memory having a defective memory cell relieving circuit
DE112007003512T5 (en) Memory device with error correction capability and efficient partial word write operation
US7076700B2 (en) Method for reconfiguring a memory
EP0892350A2 (en) Method for redundancy replacement in a memory device
TW384475B (en) Improved techniques for reducing redundant element fuses in a dynamic random access memory array
CN111128263B (en) Apparatus and method for accessing memory locations
US11776647B2 (en) Memory repair using optimized redundancy utilization
TW383485B (en) Data latch circuit for multi-level non-volatile semiconductor memory device
TW464805B (en) Distributed block redundancy for memory devices
EP0889408B1 (en) Techniques for reducing the amount of fuses in a DRAM with redundancy
JPH06187794A (en) High-efficiency dram redundant circuit
US6490209B1 (en) Memory employing multiple enable/disable modes for redundant elements and testing method using same
TW399208B (en) Improved redundant circuits and method therefor
KR20190036381A (en) Memory apparatus having SCRS and method of remapping column address in the memory apparatus

Legal Events

Date Code Title Description
GD4A Issue of patent certificate for granted invention patent
MM4A Annulment or lapse of patent due to non-payment of fees