TW200826114A - Method and apparatus for hierarchical bit line BIAS bus for block selectable memory array - Google Patents

Method and apparatus for hierarchical bit line BIAS bus for block selectable memory array Download PDF

Info

Publication number
TW200826114A
TW200826114A TW96128078A TW96128078A TW200826114A TW 200826114 A TW200826114 A TW 200826114A TW 96128078 A TW96128078 A TW 96128078A TW 96128078 A TW96128078 A TW 96128078A TW 200826114 A TW200826114 A TW 200826114A
Authority
TW
Taiwan
Prior art keywords
mode
bus
line
array
unselected
Prior art date
Application number
TW96128078A
Other languages
Chinese (zh)
Other versions
TWI345787B (en
Inventor
Roy E Scheuerlein
Luca G Fasoli
Original Assignee
Sandisk 3D Llc
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 US11/461,376 external-priority patent/US7596050B2/en
Priority claimed from US11/461,362 external-priority patent/US7633828B2/en
Application filed by Sandisk 3D Llc filed Critical Sandisk 3D Llc
Publication of TW200826114A publication Critical patent/TW200826114A/en
Application granted granted Critical
Publication of TWI345787B publication Critical patent/TWI345787B/en

Links

Classifications

    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C17/00Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards
    • G11C17/14Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards in which contents are determined by selectively establishing, breaking or modifying connecting links by permanently altering the state of coupling elements, e.g. PROM
    • G11C17/16Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards in which contents are determined by selectively establishing, breaking or modifying connecting links by permanently altering the state of coupling elements, e.g. PROM using electrically-fusible links
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C13/00Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00
    • G11C13/0002Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00 using resistive RAM [RRAM] elements
    • G11C13/0021Auxiliary circuits
    • G11C13/0023Address circuits or decoders
    • G11C13/0026Bit-line or column circuits
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C13/00Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00
    • G11C13/02Digital stores characterised by the use of storage elements not covered by groups G11C11/00, G11C23/00, or G11C25/00 using elements whose operation depends upon chemical change
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C17/00Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards
    • G11C17/14Read-only memories programmable only once; Semi-permanent stores, e.g. manually-replaceable information cards in which contents are determined by selectively establishing, breaking or modifying connecting links by permanently altering the state of coupling elements, e.g. PROM
    • G11C17/18Auxiliary circuits, e.g. for writing into memory
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C7/00Arrangements for writing information into, or reading information out from, a digital store
    • G11C7/10Input/output [I/O] data interface arrangements, e.g. I/O data control circuits, I/O data buffers
    • G11C7/1048Data bus control circuits, e.g. precharging, presetting, equalising
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C7/00Arrangements for writing information into, or reading information out from, a digital store
    • G11C7/12Bit line control circuits, e.g. drivers, boosters, pull-up circuits, pull-down circuits, precharging circuits, equalising circuits, for bit lines
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C7/00Arrangements for writing information into, or reading information out from, a digital store
    • G11C7/18Bit line organisation; Bit line lay-out
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C8/00Arrangements for selecting an address in a digital store
    • G11C8/12Group selection circuits, e.g. for memory block selection, chip selection, array selection
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C2207/00Indexing scheme relating to arrangements for writing information into, or reading information out from, a digital store
    • G11C2207/002Isolation gates, i.e. gates coupling bit lines to the sense amplifier
    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C2213/00Indexing scheme relating to G11C13/00 for features not covered by this group
    • G11C2213/70Resistive array aspects
    • G11C2213/71Three dimensional array

Landscapes

  • Engineering & Computer Science (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Semiconductor Memories (AREA)
  • Logic Circuits (AREA)
  • Read Only Memory (AREA)

Abstract

Circuits and methods are described for decoding exemplary memory arrays of programmable and, in some embodiments, re-writable passive element memory cells, which are particularly useful for extremely dense three-dimensional memory arrays having more than one memory plane. In addition, circuits and methods are described for selecting one or more array blocks of such a memory array, for selecting one or more word lines and bit lines within selected array blocks, for conveying data information to and from selected memory cells within selected array blocks, and for conveying unselected bias conditions to unselected array blocks.

Description

200826114 九、發明說明: 【發明所屬之技術領域】 本發明係關於可程式化記憶體陣列,且特定言之係關於 併入被動元件記憶體單元之半導體積體電路記憶體陣列, 且更特定言之係關於一種併入此類記憶體單元之 體陣列。 & 【先前技術】 特定被動元件記憶體單元展現可再寫特性。例如,在特 定記憶體單元中,程式化可藉由使用__大約6 w 8 v之電 壓正向偏μ記憶體單元⑽如參考其内一二極體之極性)來 實現,而抹除可藉由使用—大約10ν至14ν之電壓反向偏 屢記憶體單元來實現。該些高電壓需要在字線及位元線解 碼器内使用特殊高電塵CMOS電晶體。該些高電壓電晶體 不完全隨著記憶體單元字線及位元線間距減小而比例縮 放。此對於三維記憶體技術而言特別成問題,其中退出陣 列並必須介接一字線及位元線驅動器之字線及位元線之純 粹密度使得提供相容不斷變小陣列線間距之解碼器及ι/〇 電路(且特別係字線及位元線驅動器電路),仍能夠橫跨一 選疋5己憶體單元作用一足夠高電壓之能力丨至更加重要。 【發明内容】 般而本發明係關於一種併入一用於一區塊可選擇 記憶體陣列之階層式位元線偏壓匯流排之記憶體陣列,及 一種使用一用於一區塊可選擇記憶體陣列之階層式位元線 偏壓匯流排之方法。然而,本發明係由隨附申請專利範圍 123178.doc 200826114 來定義,故此章節内的任何内容均不應視為限制該等申請 專利範圍。 在一態樣中’本發明提供一種積體電路,其包括具有複 數個陣列區塊的-記憶體陣列,各陣列區塊包括字線與位 70線β積ϋ電路包括-般跨越該複數彳 一 一全域匯流排,用於有_合—敎區塊之敎位元線= 個別資料電路。該積體電路還包括每陣列區塊一個別第一 匯錢片段,用於在-第一操作模式期間傳遞適用於該第 -操作模式的—第—未選定位元線偏壓條件至—選定區塊 之㈣定位元線,並傳遞適用於該第一操作模式的-第二 未選定位元線偏壓條件至未選定陣列區塊之未選定位元 線。 在另1、樣中’本發明提供_種積體電路,其包括具有 複數個陣列區塊的—_悟辦 兄的A It料列,各陣列區塊包括字線盘 位元線。該積體電路包括輕合構件,其用於在一第一㈣ 杈式下糟助一般跨越該複數個陣列區塊的一第一全域匯流 排來麵0 一選定區塊之選定位元線至個別資料電路。該積 體電路包括輕合構.件,1 、 …籌件/、用於麵合選定及未選定陣列區塊 ^之€定位元線至一與各個別陣列區塊相關聯的一個 別弟一匯流排片段。&接 g “積體電路包括傳遞構件,其用於在 用…_ P相關聯之個別第-匯流排片段上傳遞適 二、„作模式的—第一未選定位元線偏屋條件。該 相闕聯之個别第—=,其用於在與該未選定陣列區塊 /机排片段上傳遞適用於該第一操作模 123l78.doc 200826114 式的一第二未選定位元線偏廢條件。 在另一態樣中,本發明提供一種用於配合一記憶體陣列 使用之方法,該記憶體陣列具有複數個陣列區塊,各陣列 區塊包括字線與位元線。該方法包括在一第一操作模式下 藉助一般跨越該複數個陣列區塊的一第一全域匯流排來耦 口選疋區塊之選定位元線至個別資料電路、耦合選定及 未選定陣列區塊二者之未選定位元線至一與各個別陣列區 塊相關聯的個別第—匯流排片段、在與該選定陣列區塊相 關聯的個別第—匯流排片段上傳_用於該第一操作模式 的第一未選定位疋線偏壓條件、及在與未選定陣列區塊 相關聯的個別第一匯流排片段上傳遞適用於該第一操作模 式之一第二未選定位元線偏壓條件。 在另一態樣中,本發明提供一種用於製造一記憶體產品 之方法。該方法包括形成一具有複數個陣列區塊之記憶體 陣列I陣列區塊包括字線與位元線。該方法還包括形 一般跨越該複數個陣列區塊的一第一全域匯流排,用於有 時麵合-選定區塊之選定位元線至個別資料電路。該方法 還包括每陣列區塊形成一個別第一匯流排片段,用於在一 第一操作模式期間傳遞適用於該第一操作模式的一第一 選定位s線偏壓條件至—選定區塊之未選定位元線 遞,該第-操作模式的一第二未選定位元線偏 至未選定陣列區塊之未選定位元線。 牛 本發明於數個態樣中適用於具有一記 路、用於操作此類積體電路及記憶體陣列之方丄= 123178.doc 200826114 入此類陣列之記憶體產品之方法及此類積體電路、產品或 記憶體陣狀電腦可讀取媒體編碼,均如本文更詳細所述 及隨附巾請專㈣圍所提出。所述技術、結構及方法可單 獨或相互組合地加以使用。 前面係一概述,因而必然包含細節之簡化、一般化及省 略因此,習知此項技術者應瞭解,前面概述僅係說明性 且不希望以任何形式限制本發明。根據下文所提出之詳細 說明’可明白僅由中請專利範圍所^義之本發明之其他態 樣、創新特徵及優點。 【實施方式】 圖1係一範例性被動元件記憶體陣列100之一示意圖。顯 示二字線102、1〇4以及二位元線106、1〇8。假定字線1〇2 係一選定字線(SWL),並假定字線1〇4係一未選定字線 (UWL)。同樣地,假定位元線1〇6係一選定位元線(sbl), 並假定位元線108係一未選定位元線(UBL)。顯示四個被動 =件記憶體單元101、103、105、1〇7,各耦合於一相關聯 字線與一相關聯位元線之間。 記憶體單元101係與選定字線1〇2及選定位元線1〇6相關 聯,故可視為一”S”單元(即”選定”單元)。記憶體單元1〇3 係與未選定字線104及選定位元線106相關聯,故可視為一 •’F”單元(即”截止”單元)。記憶體單元1〇5係與選定字線1〇2 及未選定位元線108相關聯,故可視為一,,H"單元(即,,半選 定’’單元)。最後,記憶體單元1〇7係與未選定字線1〇4及未 選疋位元線1 〇8相關聯,故可視為一,,U,,單元(即,,未選定” 123178.doc 200826114 單元)。 圖!中還說明用於-正向偏壓操作模式之範例性偏壓條 件。如本文別處所述,此正向偏壓模式可用於一程式化模 式、一區塊抹除模式及一讀取模式(但通常此類不同模式 使用不同的電麼位準或條件)。如所示,該等偏壓條件;r 視為適用於一用於一選定陣列區塊之程式化操作模式,並 將如此予以說明。 選定字線102係在一 VSX電壓(例如接地)下偏壓,選定位 元線106係在一 VSB電壓(例如+8伏特)下偏壓,未選定字線 104係在一 VUX電壓(例如+7.3伏特)下偏壓,而未選定位元 線108係在一 VUB電壓(例如+0_7伏特)下偏壓。該選定位元 線偏壓電壓VSB可視為程式化電壓νρρ,由於實質上此整 個電壓係作用於選定§己憶體單元1 〇 i上(由於該選定字線係 在接地下偏壓),在匯流排及陣列線自身内較少的特定電 阻降。該未選定位元線偏壓電壓VUB還較佳的係各記憶體 單元之一正向偏壓方向上設定在一對應於一明顯,,臨界電 壓’’之值下,因而顯示為一電壓VT正作用於未選定位元線 108上。同樣地,該未選定字線偏壓電壓νυχ還較佳的係 設定在一值VPP-VT。 在該些偏壓條件下,S單元101接收一等於VPP(例如+8 伏特)之正向偏壓電壓,F單元103接收一等於VT(例如+〇 7 伏特)之正向偏壓電壓,Η單元105接收一等於VT(例如+〇 7 伏特)之正向偏壓電壓,而U單元107接收一等於νρρ·2ντ (例如-6.6伏特)之反向偏壓電壓。存在若干範例性記憶體 123178.doc 10 200826114 單元技術,當在該些條件下偏壓時,選定單元會變化至一 更低電阻值,而該等F、Η及U單元卻絲毫沒有電阻變化。 下文說明範例性單元。 現在參考圖2,顯示用於一反向偏壓操作模式之範例性 偏壓條件200。如本文別處所述,此反向偏壓模式可用於 一程式化模式或一區塊抹除模式(但通常此類不同模式使 用不同條件)。如所示,該等偏壓條件可視為適用於一用200826114 IX. Description of the Invention: Technical Field of the Invention The present invention relates to a programmable memory array, and more particularly to a semiconductor integrated circuit memory array incorporating a passive element memory cell, and more particularly It relates to an array of bodies incorporating such memory cells. & [Prior Art] A specific passive element memory unit exhibits rewritable characteristics. For example, in a particular memory cell, stylization can be achieved by using a voltage of __about 6 w 8 v to positively bias the memory cell (10), such as referring to the polarity of a diode within it, and erasing can be performed. This is achieved by using a voltage of approximately 10 ν to 14 ν to reverse the memory cell. These high voltages require the use of special high-dust CMOS transistors in the word line and bit line decoders. The high voltage transistors are not fully scaled as the memory cell word lines and bit line pitches decrease. This is particularly problematic for three-dimensional memory technology, where the pure density of the word lines and bit lines that must exit the array and must interface with a word line and bit line drivers to provide a decoder that is compatible with ever-decreasing array line spacing. And the ι/〇 circuit (and especially the word line and bit line driver circuits) can still be more important across the ability of a selective memory cell to operate at a sufficiently high voltage. SUMMARY OF THE INVENTION The present invention generally relates to a memory array incorporating a hierarchical bit line bias bus for a block selectable memory array, and a use for one block selectable A method of a hierarchical bit line bias busbar of a memory array. However, the present invention is defined by the accompanying patent application number 123178.doc 200826114, and nothing in this section should be construed as limiting the scope of the claims. In one aspect, the present invention provides an integrated circuit including a memory array having a plurality of array blocks, each array block including a word line and a bit 70 line beta accumulation circuit including - generally spanning the complex number A global bus, used for the bit line of the _he-敎 block = individual data circuit. The integrated circuit further includes an additional first money collecting segment per array block for transmitting a first-unselected positioning element line bias condition suitable for the first operating mode during the first operating mode to - selected (4) positioning the element line of the block, and transmitting a second unselected positioning element line bias condition suitable for the first mode of operation to an unselected positioning element line of the unselected array block. In another example, the present invention provides an integrated circuit including an A It column having a plurality of array blocks, each array block including a word line disk bit line. The integrated circuit includes a light-fitting component for facilitating a first global busbar that generally spans the plurality of array blocks in a first (four) mode, and selects a selected location line from the selected block to Individual data circuits. The integrated circuit includes a light-construction component, a ..., a component, a positioning element for the face-selecting and unselected array blocks, and a different one associated with each of the array blocks. Bus segment. & g "The integrated circuit includes a transfer member for transmitting the appropriate condition of the second unselected positioning element on the individual first-bus bar segment associated with ..._P. The individual number -= of the phase combination is used to pass a second unselected positioning element line depletion condition applicable to the first operating mode 123l78.doc 200826114 on the unselected array block/machine segment. In another aspect, the invention provides a method for use with a memory array having a plurality of array blocks, each array block comprising word lines and bit lines. The method includes, in a first mode of operation, coupling a selected location line to a plurality of data circuits, coupling selected and unselected array regions by means of a first global bus bar that generally spans the plurality of array blocks An unselected positioning element line of the block to an individual first bus bar segment associated with each of the individual array blocks, and an individual first bus bar segment associated with the selected array block uploading_for the first Transmitting a first unselected positioning line bias condition of the operating mode and transmitting an individual unselected positioning line offset for the first operating mode on the individual first bus segment associated with the unselected array block Pressure conditions. In another aspect, the invention provides a method for making a memory product. The method includes forming a memory array I array block having a plurality of array blocks including word lines and bit lines. The method also includes forming a first global busbar that generally spans the plurality of array blocks for use in selecting the location lines of the time-selected-selected blocks to the individual data circuits. The method also includes forming a different first busbar segment per array block for transmitting a first selected positioning s-wire bias condition suitable for the first mode of operation to a selected block during a first mode of operation The unselected positioning element line, the second unselected positioning element line of the first operation mode is biased to the unselected positioning element line of the unselected array block. The invention of the invention is applicable to a method for operating such an integrated circuit and a memory array in a plurality of aspects: 123178.doc 200826114 A method for entering a memory product of such an array and such a product The body code, product or memory array computer can read the media code, as described in more detail in this article and attached to the towel (4). The techniques, structures and methods can be used individually or in combination with one another. The foregoing is a summary of the invention, and is therefore intended to be in the Other aspects, innovative features, and advantages of the invention will be apparent from the scope of the appended claims. [Embodiment] FIG. 1 is a schematic diagram of an exemplary passive component memory array 100. Two word lines 102, 1 〇 4 and two bit lines 106, 1 〇 8 are displayed. Assume that word line 1 〇 2 is a selected word line (SWL) and that word line 1 〇 4 is an unselected word line (UWL). Similarly, the pseudo-positioning line 1〇6 is a selected positioning element line (sbl), and the pseudo-positioning unit line 108 is an unselected positioning element line (UBL). Four passive = piece memory cells 101, 103, 105, 1 〇 7 are shown, each coupled between an associated word line and an associated bit line. The memory unit 101 is associated with the selected word line 〇2 and the selected locating element line 〇6, and thus can be regarded as an "S" unit (i.e., "selected" unit). The memory unit 1〇3 is associated with the unselected word line 104 and the selected positioning element line 106, so it can be regarded as a 'F' unit (ie, "off" unit). The memory unit 1〇5 is connected to the selected word line. 1〇2 and unselected positioning element line 108 are associated, so it can be regarded as one, H" unit (ie, semi-selected ''unit). Finally, memory unit 1〇7 is connected with unselected word line 1〇4 And the unselected bit line 1 〇 8 is associated, so it can be regarded as one, U,, unit (ie, not selected) 123178.doc 200826114 unit). Figure! Exemplary bias conditions for the -forward bias mode of operation are also described. As described elsewhere herein, this forward bias mode can be used in a stylized mode, a block erase mode, and a read mode (but typically such different modes use different levels or conditions). As shown, these bias conditions; r are considered to be suitable for a stylized mode of operation for a selected array block and will be described as such. The selected word line 102 is biased at a VSX voltage (e.g., ground), the selected bit line 106 is biased at a VSB voltage (e.g., +8 volts), and the unselected word line 104 is tied to a VUX voltage (e.g., + The 7.3 volts are biased, while the unselected location lines 108 are biased at a VUB voltage (e.g., +0-7 volts). The selected positioning element line bias voltage VSB can be regarded as a stylized voltage νρρ, since substantially the entire voltage is applied to the selected § memory unit 1 〇i (since the selected word line is biased at ground), There is less specific resistance drop in the bus and array lines themselves. The unselected positioning element line bias voltage VUB is also preferably set in a forward bias direction of one of the memory cells to correspond to an apparent, threshold voltage '' value, thus being displayed as a voltage VT Acting on the unselected positioning element line 108. Similarly, the unselected word line bias voltage ν υχ is preferably set to a value of VPP-VT. Under these bias conditions, S unit 101 receives a forward bias voltage equal to VPP (e.g., +8 volts) and F unit 103 receives a forward bias voltage equal to VT (e.g., +〇7 volts). Unit 105 receives a forward bias voltage equal to VT (e.g., +〇7 volts), and U unit 107 receives a reverse bias voltage equal to νρρ·2ντ (e.g., -6.6 volts). There are several exemplary memories. 123178.doc 10 200826114 Unit technology, when biased under these conditions, the selected cells will change to a lower resistance value, while the F, Η and U cells have no resistance change. Exemplary units are described below. Referring now to Figure 2, an exemplary bias condition 200 for a reverse bias mode of operation is shown. As described elsewhere herein, this reverse bias mode can be used in either a stylized mode or a block erase mode (although typically different modes use different conditions). As shown, these bias conditions can be considered to be suitable for one use.

於一選定陣列區塊之程式化操作模式或抹除操作模式,並 將如此予以說明。 該等偏壓條件VSX、Vux、VSB及VUB之各偏壓條件現 在將針對適用於本操作模式之值來加以重新定義。選定字 線102係在一 Vsx電壓VRR/2(例如+5伏特)下偏壓,而選定 位兀線106則在一 VSB電壓-VRR/2(例如_5伏特)下偏壓。該 未選疋予線電壓VUX及該未選定位元線電壓vub二者均接 地0 ,嚙、 ▼ Αν XV1 jyfi 在該些偏壓條件下,S單元ιοί接收一數 如-ίο伏特)之反向偏壓電壓,F單元1〇3接收一等於 ^ΚΚ/2(例如·5伏特)之反向偏壓電壓,而Η單元1G5接收一 等於VRR/2(例如_5伏特)之反向偏壓電壓。應注意 1〇7橫跨單元上不接收任何偏壓。 子在右干範例性記憶體單元技術(參考下面),當在該些 條件下偏壓時,選 一 &疋早兀會從一較低電阻值變化至一更言 電阻值,而綠回 立 q u單元電阻卻沒有絲毫變化。還應 注思’該等去遁中 、& u兒憶體單元沒有任何偏壓,因而沒有 123178.doc • 11 · 200826114 任何漏電流,其在橫跨此類單元在數伏特下偏壓時另外可 月b支杈一相當大數量的洩漏電流。如進一步詳細所述,許 多有用記憶體陣列具體實施例包括比H單元或F單元遠大得 多數目的U單70,且此類陣列較其他偏壓方案在陣列之該 等未選定記憶體單元内具有明顯較低的漏電流以及因此低 得多的功率消耗。 藉由在此反向模式下,,分割” VRR電壓,並在一等於程式 化電壓-半的負電壓下偏壓飢,以及在—等於程式化電 壓-半的正f壓下偏壓SWL,位元線解碼器及字線解碼器 者之電壓要求得到明顯鬆弛。因此,與該等陣列線(例 ^字線及位TL線)之較小間距相一致,在該等陣列線驅動 器電路内的該等高電壓電晶體佔據更少㈣,因為其可設 計以獲得一相對較低的”分割”電壓。 其他記憶體技術-直面對關於程式化及抹除電壓(及此 類高電壓電晶體所需之面積)不隨記憶體單元間距以相同 速率比例縮放之類似問題。例如,在快閃記憶體内此問題 的影響有時因為-般以快閃記憶體為主記憶體陣列之更大 扇出而稍微降低。用於高電壓電晶體之更多^肖耗設計 規則可藉由增加記憶體區塊大小而在某些更新穎技術中得 •肖r、、、:而在以一極體為主被動元件記憶體陣列 中’更大區塊大小細透過屬於選定陣狀該等未選定記 :體單元之㈣漏為代價的。#由如圖2所示偏壓此類 〜‘定。己|·思體單兀’可將此洩漏成分減小至幾乎零,並獲 得—更大區塊大小’同時幾乎沒有有害功率消耗。 " 123178.doc 200826114 現在參考圖3,顯示一範例性字線解碼器電路,包括顯 示適用於正向偏壓操作模式之偏壓條件(如圖1所示)。在頁 面左側顯示一列解碼器電路,其顯示二解碼輸出15 8、 162。解碼輸出158對應於一選定解碼輸出,而解碼輸出 162對應於一未選定解碼輸出。一列解碼器152可使用各種 熟知技術之任一者來實施,產生複數個解碼輸出,例如輸 出155、159,其係由多工器157、161及反相器156、16〇來 有條件地反轉。一反轉緩衝器係在該NAND閘之後併入以 由於在節點158上的較大電容性負載而驅動節點155(即結 果,如此處,多工器157操縱節點155至輸出158)。列解碼 器152係在此操作模式下操作…等於vpp之上部供應電 壓係輕合至電源節點153以及一接地下部供應電壓係輕合 至電源節點154。在此操作模式下,該列解碼器係一 ”高態 有效解碼裔,意味著將諸如解碼輸出節點158之選定輸出 (或多個輸出)驅動至兩個可用電壓狀態之最高者,在此情 況下係憎。該等未選定解碼輸出(例如解碼輸出節點叫 係驅動至該等兩個可用電壓狀態之最低者,在此情況下係 接地了述將最魏定—次僅選擇—此類解碼輸點 (例如”高態”)。 各解碼輸出_合至-❹個字線驅動器電路。例如, 解碼輸出節點⑸係耗合至一字線驅動器電路括 :刪電晶體171與卿s電晶體m。電晶體171:、172之 個別汲極端子係同時耦合至一 字線心儘管本發明之特定::眚在此情況下表示選定 之特疋具體實施例涵蓋多頭解碼器 123178.doc •13- 200826114 外的解碼器,但圖3描述一還耦合至解碼輸出節點i58之第 二字線驅動器電路,其表示與此特定解碼輸出節點158相 關聯之一或多個其他字線驅動器電路。此第二字線驅動器 電路包括PMOS電晶體173及NMOS電晶體174,其輸出驅 動一字線1 8 1,字線1 8 1表示一或多個半選定字線。 在該些字線驅動器電路之各電路内的NM0S電晶體之個 別源極端子係耦合至一源極選擇匯流排XSEL之一個別匯 流排線。在此操作模式下,該源極選擇匯流排係基於位址 資訊來解碼’使得在一適用於此操作模式一字線之有效狀 ; 態下偏壓此類匯流排線,而在一適用於此操作模式字線之 無效操作模式下偏壓剩餘匯流排線。在特定具體實施例 中’多個此類源極選擇匯流排可以係有效的,但現在假定 匯流排線167有效,並在接地下偏壓,而一或多個剩餘匯 流排線(表示為匯流排線1 68)係無效並驅動至該未選定字線 電壓VUX(顯示為VPP-VT)。 由於在解碼輸出節點158上的電壓(VPP)高於匯流排線 167、168之電壓,故該等NMOS電晶體172、174二者係接 通,從而驅動選定字線102接地,並驅動半選定字線181至 VPP-VT。該些二傳導路徑係指示為開尾式箭頭線。 在該些字線驅動器電路之各電路内的PM〇s電晶體之個 別源極端子係耦合至一未選定偏壓線UXL,還標註為節點 1 64。在此操作模式下,該UXL偏壓線傳遞該未選定字線 電壓VUX。由於在解碼輸出節點158上的電壓(vpp)高於該 UXL偏壓線之電壓,故二pm〇S電晶體171、173係截止。 123178.doc -14- 200826114 解碼輸出即點162係耦合至一字線驅動器電路,其包括 PMOS電晶體175與職〇8電晶體。電晶體、μ之 個別汲極端子係同時耦合至一字線,在此情況下表示未選 疋字線104。還耦合至解碼輸出節點162之一第二字線驅動 裔電路表示與解碼輸出節點162相關聯之一或多個剩餘字 線驅動器電路,並包括PMOS電晶體177及NMOS電晶體 178,該等電晶體之輸出驅動一未選定字線183。 如如述’在該些字線驅動器電路之各電路内的Nm〇s電 晶體之個別源極端子係耦合至一源極選擇匯流排xsel之 一個別匯流排線。由於在解碼輸出節點162上的電壓(接地) 係在或低於匯流排線167 ' 168之電壓,故該等NM〇s 、178二者係截止。在該些字線驅動器電路之各電路内 的PMOS電晶體之個別源極端子係耦合至未選定偏壓線 UXL節點164。由於在解碼輸出節點162上的電壓(接地)係 低於UXL偏壓線164之電壓(低PMOS臨界電壓以上),二 PMOS電晶體175、177係接通,從而驅動該等未選定字線 104、1 83至VUX(例如VPP-VT)。該些二傳導路徑係指示為 開尾式箭頭線。 現在參考圖4,顯示此相同範例性字線解碼器電路,包 括適用於反向偏壓操作模式之偏壓條件(如圖2所示)。該列 解碼電路之解碼輸出158仍對應於一選定解碼輸出,而解 碼輸出162對應於一未選定解碼輸出。列解碼器ι52係在此 操作模式下操作,一等於VRR/2之上部供應電壓係輕合至 電源節點153以及一接地下部供應電壓係耦合至電源節點 123178.doc 200826114 154。在此操作模式下,該列解碼器係一,,高態有效,,解碼 器’而該有效(選定)解碼輸出158係使用反相器156及多工 器1 57而驅動至兩個可用電壓狀態之最低者,在此情況下 其係GND(接地)。該等未選定解碼輸出(例如解碼輸出節點 162)係使用反相器16〇及多工器161而驅動至該等兩個可用 電壓狀態之最高者,在此情況下係VRR/2。 在此操作模式下,對於所述範例性具體實施例而言,源 極選擇匯流排XSEL之該等個別匯流排線係全部驅動至相 同偏壓條件(接地),而該”未選定,,偏壓線11又1^傳遞一等於 VRR/2(例如+5伏特)之偏壓電壓Q在此反向操作模式下, 該偏壓線UXL實際上傳遞一適用於字線之有效狀態,而非 一無效或未選定偏壓條件。由於在解碼輸出節點158上的 電壓(GND)相當低於該偏壓線UXL之電壓(即低一 pM〇s臨 界電壓以上),該等PM0S電晶體171、173二者係接通,從 而將該選定字線102驅動至VRR/2,並還將本來係半選擇 字線者(此處顯示為選定字線181)驅動至VRR/2。該些二傳 導路徑係指示為開尾式箭頭線。 在此操作模式下,不解碼該源極選擇匯流排xsel,且 各此類匯流排線係在一適用於一字線之無效狀態下(例如 接地)偏壓。由於在解碼輸出節點158上的電壓(接地)不高 於匯流排線167、168之電壓,故該等1^1^〇8 172、174二者 係截止。 解碼輸出節點162係一未選定輸出,藉由反相器16〇及多 工器161而驅動至VRR/2。由於在解碼輸出節點丨62上的電 123178.doc -16- 200826114 壓尚於匯流排線167、168之電壓,故該等NMOS 176、178 一者係接通’從而將該等未選定字線丨〇4、ί 83驅動至接 地0亥些一傳^路徑係指示為開尾式箭頭線。由於在解碼 輸出節點162上的電壓與該UXL偏壓線164上傳遞的電壓相 同,故二PMOS電晶體175、177係截止。 ίThe stylized mode of operation or the erase mode of operation of a selected array block will be described as such. The bias conditions of these bias conditions VSX, Vux, VSB and VUB will now be redefined for values suitable for this mode of operation. The selected word line 102 is biased at a Vsx voltage of VRR/2 (e.g., +5 volts) and the selected bit line 106 is biased at a VSB voltage of -VRR/2 (e.g., _5 volts). The unselected sputum line voltage VUX and the unselected locating element line voltage vub are both grounded to 0, bite, ▼ Αν XV1 jyfi under the bias conditions, S unit ιοί receives a number such as - ίο volts To the bias voltage, F unit 1〇3 receives a reverse bias voltage equal to ^ΚΚ/2 (eg, 5 volts), and unit 1G5 receives a reverse bias equal to VRR/2 (eg, _5 volts). Voltage. It should be noted that 1〇7 does not receive any bias voltage across the unit. In the right-handed exemplary memory cell technology (see below), when biased under these conditions, the selection of a & early will change from a lower resistance value to a more specific resistance value, while green back There is no slight change in the resistance of the vertical qu unit. It should also be noted that 'there are no biases in the body, and there is no leakage current, which is any voltage leakage across several volts across such cells. In addition, a considerable amount of leakage current can be supported by the monthly b. As described in further detail, many useful memory array embodiments include a much larger number of U-singles 70 than H cells or F cells, and such arrays have in the unselected memory cells of the array compared to other biasing schemes Significantly lower leakage current and therefore much lower power consumption. By splitting the "VRR voltage" in this reverse mode and biasing the hunger at a negative voltage equal to the programmed voltage - half, and biasing SWL at a positive f-voltage equal to the programmed voltage - half, The voltage requirements of the bit line decoder and the word line decoder are significantly relaxed. Therefore, consistent with the smaller spacing of the array lines (eg, word lines and bit TL lines), within the array line driver circuits The high voltage transistors occupy less (4) because they can be designed to achieve a relatively low "divided" voltage. Other memory technologies - straightforward about stylized and erase voltages (and such high voltages) The area required for the crystal does not scale at the same rate as the memory cell spacing. For example, the effect of this problem in flash memory is sometimes due to the fact that flash memory is the main memory array. Large fanouts are slightly reduced. More design rules for high-voltage transistors can be achieved by increasing the memory block size in some newer technologies. Polar body-based passive component memory array The size of the larger block is at the cost of the unselected note belonging to the selected array: (4) the leakage of the body unit. ################# 'This leakage component can be reduced to almost zero and obtained - larger block size' with almost no harmful power consumption. " 123178.doc 200826114 Referring now to Figure 3, an exemplary word line decoder circuit is shown, including A bias condition suitable for the forward bias mode of operation is shown (as shown in Figure 1). A column of decoder circuits is shown on the left side of the page that displays two decoded outputs 15 8 and 162. The decoded output 158 corresponds to a selected decoded output. The decoded output 162 corresponds to an unselected decoded output. A column of decoders 152 can be implemented using any of a variety of well known techniques to produce a plurality of decoded outputs, such as outputs 155, 159, which are multiplexers 157, 161 and Inverters 156, 16 are conditionally inverted. An inverting buffer is incorporated after the NAND gate to drive node 155 due to the large capacitive load on node 158 (i.e., as herein, Multiplexer 157 manipulation Node 155 to output 158). Column decoder 152 operates in this mode of operation ... equal to vpp upper supply voltage is coupled to power supply node 153 and a grounded lower supply voltage is coupled to power supply node 154. In this mode of operation Next, the column decoder is a "high state efficient decoding", meaning that the selected output (or outputs), such as the decoded output node 158, is driven to the highest of the two available voltage states, in which case it is tied. The unselected decoded outputs (e.g., the decoded output node is driven to the lowest of the two available voltage states, in which case the grounded pair will be most determined - only selected - such decoded input points (eg "High state") Each decoding output _ is connected to - one word line driver circuit. For example, the decoding output node (5) is consuming to a word line driver circuit including: a transistor 171 and a transistor s. 171: 172, the individual 汲 extremes are simultaneously coupled to a word line heart. Although the invention is specific: 眚 in this case indicates that the selected embodiment covers the multi-head decoder 123178.doc •13-200826114 The decoder, but Figure 3 depicts a second word line driver circuit also coupled to the decoded output node i58, which represents one or more other word line driver circuits associated with this particular decoded output node 158. This second word line The driver circuit includes a PMOS transistor 173 and an NMOS transistor 174, the output of which drives a word line 181, and the word line 181 represents one or more semi-selected word lines. Within each circuit of the word line driver circuits NM0S electric crystal The individual source terminal is coupled to an individual bus line of one of the source select busses XSEL. In this mode of operation, the source select bus is decoded based on the address information to make it suitable for this mode of operation. The word line is active; the bus bar is biased in the state, and the remaining bus bars are biased in an inactive mode of operation suitable for the operating mode word line. In a particular embodiment, 'a plurality of such The source select bus can be active, but now assume that the bus bar 167 is active and biased at ground, and one or more of the remaining bus bars (denoted as bus bar 168) are invalid and driven to The word line voltage VUX (shown as VPP-VT) is not selected. Since the voltage (VPP) on the decoded output node 158 is higher than the voltage of the bus bars 167, 168, the NMOS transistors 172, 174 are connected. Passing, thereby driving the selected word line 102 to ground, and driving the semi-selected word lines 181 to VPP-VT. The two conductive paths are indicated as open-ended arrow lines. PM〇 in each circuit of the word line driver circuits s individual source terminal of the transistor The unselected bias line UXL is also labeled as node 1 64. In this mode of operation, the UXL bias line passes the unselected word line voltage VUX. Because the voltage (vpp) at the decoded output node 158 is high. The voltage of the UXL bias line is such that the two pm 〇S transistors 171, 173 are turned off. 123178.doc -14- 200826114 The decoded output, point 162, is coupled to a word line driver circuit, which includes a PMOS transistor 175 and The operating transistor 8 transistor, the individual terminal of the transistor, is simultaneously coupled to a word line, in which case the unselected word line 104 is shown. Also coupled to one of the decoded output nodes 162, the second word line driver circuit represents one or more remaining word line driver circuits associated with the decoded output node 162 and includes a PMOS transistor 177 and an NMOS transistor 178, such The output of the crystal drives an unselected word line 183. The individual source terminals of the Nm〇s transistors in each of the circuits of the word line driver circuits are coupled to an other bus bar of a source select bus xsel. Since the voltage (ground) at the decode output node 162 is at or below the voltage of the bus bar 167 ' 168, both NM 〇 s , 178 are off. The individual source terminals of the PMOS transistors in the various circuits of the word line driver circuits are coupled to unselected bias line UXL node 164. Since the voltage (ground) at the decode output node 162 is lower than the voltage of the UXL bias line 164 (above the low PMOS threshold voltage), the two PMOS transistors 175, 177 are turned "on", thereby driving the unselected word lines 104. , 1 83 to VUX (for example, VPP-VT). The two conduction paths are indicated by an open-ended arrow line. Referring now to Figure 4, this same exemplary word line decoder circuit is shown, including bias conditions suitable for the reverse bias mode of operation (shown in Figure 2). The decoded output 158 of the column decode circuit still corresponds to a selected decoded output, and the decoded output 162 corresponds to an unselected decoded output. The column decoder ι52 operates in this mode of operation, with a supply voltage equal to the upper portion of VRR/2 coupled to the power supply node 153 and a lower ground supply voltage coupled to the power supply node 123178.doc 200826114 154. In this mode of operation, the column decoder is one, the high state is active, and the decoder 'and the active (selected) decoded output 158 is driven to two available voltages using inverter 156 and multiplexer 157. The lowest of the states, in this case it is GND (ground). The unselected decoded outputs (e. g., decoded output node 162) are driven to the highest of the two available voltage states using inverter 16 and multiplexer 161, in this case VRR/2. In this mode of operation, for the exemplary embodiment, the individual bus bars of the source select bus XSEL are all driven to the same bias condition (ground), and the "unselected, biased" The voltage line 11 further transmits a bias voltage Q equal to VRR/2 (e.g., +5 volts). In this reverse mode of operation, the bias line UXL actually delivers an active state suitable for the word line, instead of An invalid or unselected bias condition. Since the voltage (GND) on the decoded output node 158 is relatively lower than the voltage of the bias line UXL (i.e., above a pM s s threshold voltage), the PMOS transistors 171, 173 is both turned "on" to drive the selected word line 102 to VRR/2, and also drives the originally half-selected word line (shown here as selected word line 181) to VRR/2. The path is indicated by an open-ended arrow line. In this mode of operation, the source select bus xsel is not decoded, and each such bus line is in an inactive state (eg, ground) that is applied to a word line. Voltage because the voltage (ground) at the decoded output node 158 is not higher than The voltages of the bus bars 167, 168 are such that the 1^1^8 172, 174 are both turned off. The decoded output node 162 is an unselected output, driven by the inverter 16 and the multiplexer 161. Up to VRR/2. Since the voltage 123178.doc -16-200826114 on the output node 丨62 is pressed against the voltage of the bus bars 167, 168, the NMOS 176, 178 is turned "on" The unselected word line 丨〇4, ί 83 is driven to ground 0. The path is indicated as an open-ended arrow line. Due to the voltage on the decoded output node 162 and the voltage transmitted on the UXL bias line 164 The same, so the two PMOS transistors 175, 177 are cut off.

現在參考圖5,顯示一範例性位元線解碼器電路,包括 顯示適用於正向偏壓操作模式之偏壓條件(如圖1所示)。在 頁面左側顯示一行解碼器電路,其顯示二解碼輸出2〇8、 212。解碼輸出208對應於一選定解碼輸出,而解碼輸出 212對應於一未選定解碼輸出。一行解碼器2〇2可使用各種 熟知技術之任一者來實施,產生複數個解碼輸出,例如輸 出205、209,其係由多工器207、211及該等反相器2〇6、 210來有條件地加以反轉。不同於該列解碼器,在該^八^〇 閘之後不存在反轉緩衝器以驅動節點2〇5,因為在節點2〇8 上的電容性負載要比用於該等列解碼器輸出的要低得多。 行解碼器202係在此操作模式下操作,—等於之上呷 供應電壓係耦合至電源節點203以及一接地下部供應電壓 係麵合至電源節點204。在此操作模式下,該行解碼器係 低態有/效&quot;解碼器。該等未選定解碼輸出(例如解碼輸出 節點212)係驅動至該等兩個可用電壓狀態之最高者,在此 情況下係VPP。下述將最初假定一次僅選 &lt;俾此類解碼輸 出節點208(例如”低態”)。 各解碼輸出係耦合至-或多個位元線驅動器電路。例 如,解碼輸出節點208係耦合至一位元線驅動器電路,其 123178.doc 200826114 包括PMOS電晶體221與NMOS電晶體222。電晶體221、 222之個別汲極端子係同時耦合至一位元線,在此情況下 表示選定位元線106。儘管本發明之特定具體實施例涵蓋 多頭解碼器外的解碼器,但圖5描述一還輕合至解碼輸出 節點208之第二位元線驅動器電路,其表示與此特定解碼 輸出節點208相關聯之一或多個剩餘位元線驅動器電路。 此第二字線驅動器電路包括PM〇s電晶體223&amp;nm〇s電晶 體224 β亥荨電晶體之輸出驅動一位元線23 1,其表示一戋 多個半選定位it、線。比較該字線解碼器,&amp;類半選定位元 線可表示一選定位元線,其係維持在一無效狀態下。 在忒些位元線驅動器電路之各電路内的pM〇s電晶體之 個別源極端子係耦合至一源極選擇匯流排selB2一個別 匯流排線。在此操作模式下,該源極選擇匯流排卯⑶係 資料相依的,且可進一步基於位址資訊來加以解碼,使得 對於此操作模式,一或多個此類匯流排線係在一適用於一 位元線之有效狀態下偏壓,而對於此操作模式,剩餘匯流 排線係在一適用於位元線之無效狀態下偏壓。在特定具體 實施例中,一或多個此類源極選擇匯流排線可以係有效 的,但現在假定匯流排線217係有效,並在vpp下偏壓, 而一或多個剩餘匯流排線(表示為匯流排線2丨係無效並驅 動至該未選定位元線電壓VUB(顯示為VT)。 由於在解碼輸出節點208上的電壓(接地)低於匯流排線 217、218之電壓,故該等?]^〇8電晶體221、223二者係接 通,從而驅動選定位元線106至VPP,並驅動半選定位元 123178.doc -18- 200826114 線23 1至VT。該些二傳導路徑係指示為開尾式箭頭線。 在該些位元線驅動器電路之各電路内的NMOS電晶體之 個別源極端子係耦合至一未選定偏壓線UYL,還標註為節 點214。在此操作模式下,該UYL偏壓線傳遞該未選定位 元線電壓VUB。由於在解碼輸出節點208上的電壓(接地)低 於該UYL偏壓線之電壓,故二NMOS電晶體222、224係截 止。 解碼輸出節點212係耦合至一位元線驅動器電路,其包 括PMOS電晶體225與NMOS電晶體226。電晶體225、226 之個別汲極端子係同時耦合至一位元線,在此情況下表示 未選定位元線108。還耦合至解碼輸出節點212的一第二位 元線驅動器電路表示與解碼輸出節點212相關聯的一或多 個剩餘位元線驅動器電路,並包括PMOS電晶體227及 NMOS電晶體228,該等電晶體之輸出驅動一未選定位元線 233 ° 如上述,在該些位元線驅動器電路之各電路内的PMOS 電晶體之個別源極端子係耦合至一源極選擇匯流排SELB 之一個別匯流排線。由於在解碼輸出節點212上的電壓 (VPP)係在或高於匯流排線217、218之電壓,故該等PMOS 225、227二者係截止。在該些位元線驅動器電路之各電路 内的NMOS電晶體之個別源極端子係耦合至未選定偏壓線 UYL節點214。由於在解碼輸出節點212上的電壓係VPP, 故二NMOS電晶體226、228係接通,從而將未選定位元線 108、233驅動至VUB(例如VT)。該些二傳導路徑係指示為 123178.doc -19- 200826114 開尾式箭頭線。 現在參考圖6,顯示該位元線解碼器電路,包括適用於 反向偏塵操作模式之偏壓條件(如圖2所示)。該行解碼器電 路之解碼輸出208仍對應於一選定解碼輸出,而解碼輸出 2 12對應於一未選定解碼輸出。行解碼器2〇2係在此操作模 式下操作,一等於GND之上部供應電壓係耦合至電源節點 203以及一下部供應電壓_Vrr/2係耦合至電源節點2〇4。在 此操作模式下,該行解碼器係一&quot;高態有效”解碼器,而該 有效(選定)解碼輸出208係藉由反相器206及多工器207而驅 動至兩個可用電壓狀態之最高者,在此情況下其係 GND(接地)。該等未選定解碼輸出(例如解碼輸出節點212) 係藉由反相器210及多工器211而驅動至該等兩個可用電壓 狀態之最低者,在此情況下其係—VRR/2。 在此操作模式下,對於所述範例性具體實施例而言,源 極選擇匯流排SELB之該等個別匯流排線係全部驅動至相 同偏壓條件(接地),而該”未選定&quot;偏壓線UYL傳遞一等 於-VRR/2(例如-5伏特)之偏壓電壓。在此反向操作模式 下,該偏壓線UYL實際上傳遞一適用於位元線之有效狀 態,而非一無效或未選定偏壓條件。由於在解碼輸出節點 208上的電壓(接地)相當程度高於該偏壓線uyL之電壓(低 一 NMOS臨界電壓以上),該等NMOS電晶體222、224二者 係接通’從而將該選定位元線106驅動至-VRR/2,並還將 本來係半選擇位元線者(此處顯示為選定位元線23丨)驅動 至-VRR/2。該些二傳導路徑係指示為開尾式箭頭線。 123178.doc -20- 200826114 在此操作模式下,該源極選擇匯流排SELB係非資料相 依或不解碼(至少在-給定區塊内),且各此類匯流排線係 在一適用於一位元線之無效狀態下(例如接地)偏壓。該等 PMOS電晶體221、223二者係截止。 解碼輸出節點212係一未選定輸出並驅動至-VRR/2。該 PMOS電晶體225、227二者係接通,從而驅動該等未選 疋位兀線108、233至接地。該些二傳導路徑係指示為開尾 式箭頭線。二NMOS電晶體220、228係截止。 々應注意’在該正向模式下’該行解碼器係低態有效而該 等位το線係高態有效。但在該反向模式下,該行解碼器逆 反其極性而變成高態有效,而該等位元線自身也逆反極性 而變成低態有效。反之’在該正向模式下,該列解碼器係 高態有效而該等字線係低態有效。但在該反向模式下,該 列解碼H逆反其極性而變成低態有效,而該等字線自身也 逆反極性而變成高態有效。還應注意,該行解碼器輸出位 準在該正向模式(即GND至VPP)與反向模式(即_VRR/2至 GND)之間在平均電壓上偏移。 當視為一非多頭解碼器(在圖3、4、5及6中,僅非虛線 陣列線驅動器電路)時,可極簡單地說明該解碼器電路之 操作。在該反向模式下,該字線解碼器逆反其極性並使一 選定字線成高態(〜5 V),同時保持所有其他選定字線接 地。該逆反發生於位域選擇側,其中敎—位元線並成 為-5 V而所有其他位元線均接地。最終結果係橫跨選定記 憶體單元的10 V反向偏壓與橫跨其他單元的零反向偏壓。 123178.doc -21 - 200826114 等子線及位元線驅動II電路内的該等電晶體僅須承受 5 V,或最大電壓一半,而非整個電壓。 當考量使用多頭解石馬器(在圖3、4、5及6中,包括虛陣 列線驅動器電路)之薇涵時,應注意,至此所述電路在該 正向方向上利用-解碼源極選擇匯流排,其允許選擇該陣 列線群組之-單一者(但是同時將剩餘半選定陣列線驅動 至一未選定㈣條件)。然而,在該反向模式下,來自該 列及行解碼器之選定解碼輸出將各陣列線耦合至一單一未 選^偏壓線(例如UXL及UYL)。使用m線在該反 向权式下無法獲知半選定陣列線。由此,在配置用以在該 反向模式下選擇一陣列線區塊(例如一,,區塊抹除,,)時上述 電路及技術非常有用。如在圖4及6中可看出,一選定字線 區塊與一選定位元線區塊在該反向模式下同時選擇,沒有 任何獨立可組態的半選定陣列線。此類區塊操作全部避免 任何半選定線之需要。解碼蘊涵可能極類似於授予ε· &gt;heUerlein之美國專利第6,879,5〇5號,標題為&quot;用於三維 記憶體陣列之具有多層字線片段之字線配置,,中所述,其 揭示内谷王。卩以引用形式併入本文。是否可組態此類區塊 操作(或可組態多大的區塊)主要取決於單元重置電流之數 量、同時傳導此類重置電流之單元數目、以及在字線驅動 器電路及位元線驅動電路内的PM〇s與NMOS電晶體是否 可在可接受電壓降下支援此類電流。 可藉由使用其他技術在該反向模式下提供半選定陣列線 (除了在该正向模式下已經提供的)。在一單一此類技術 123178.doc -22- 200826114 中,該等列及行解碼器可由過電壓來供電,使得該等解碼 輸出節點高於該PMOS源極電壓且低於該NM0S源極電壓 而橫過。藉由如此操作,可透過NM〇s電晶體將選定字線 驅動最南至+VRR/2電壓,並可透過PM〇s電晶體將選定位 兀線驅動最低至-VRR/2電壓。此利用與該正向模式期間相 同的電晶體來驅動選定字線及位元線。 此類技術如圖7及8所示。最初參考圖7,說明一字線解 碼為電路,其利用一過驅動解碼輸出來驅動該等陣列線驅 動器,其源極保持在上述偏壓條件下。在此列解碼器電路 中,列解碼器152係由一8伏特上部供應電壓與一負i伏特 下部供應電壓來供電。該等解碼輸出節點158、162之極性 ,相對於圖4所示之極性逆反,故現在係一高態有效解碼 器,其在+8伏特下提供一選定輸出158並在_丨伏特下提供 一未選定解碼輸出162。源極選擇匯流排xsel保持一解碼 匯流排不變。其個龍流排線之―(或多個)者係選定並驅 動至+5伏特,而該等未選定線係驅動至接地。nm〇s電晶 體172係接通,並將選定字線1〇2傳導至相關聯的職匯 流排線電壓(+5伏特)。NM〇s電晶體m係也接通,並將該 (等)半選定字線181傳導 1寻導至接地。在未選定解碼輸出節點 叫-&quot;犬特下時’該等PMOS電晶體175、177係同時接 通,並將未選W線πμ、183傳導至接地。在利用此技術 之某些具體實施射,不使㈣等條件輸出反相器156、 160及該等多工器157、161(此處顯示為&quot;虛線”)。 現在參考圖8,說明一位元線解碼器電路,其也利用一 123178.doc -23 - 200826114 過驅動解碼輸出來驅動該等陣列線驅動器。在此行解碼器 電路中,行解碼器202係由一 +1伏特高電源電壓與一負8伏 特低電源電壓來供電。該等解碼輸出節點2〇8、212之極性 係相對於圖6所示之極性逆反,故現在係一低態有效解碼 器,其在_8伏特下提供一選定輸出208並在+1伏特下提供 一未選定解碼輸出212。該等個別8£1^6匯流排線217之一 (或多個)者係選定並驅動至_5伏特,而該等未選定犯16匯 流排線218係驅動至接地。PM〇s電晶體221係接通,並將 名等選疋位元線1 〇6傳導至相關聯的SELB匯流排線電壓(_5 伏特)。PMOS電晶體223係也接通,並將該(等)半選定字線 23 1傳導至接地。在未選定解碼輸出節點212在+1伏特下 時,忒# NMOS電晶體226、228係同時接通,並將該等未 選定位元線108、233傳導至接地。在利用此技術之某些具 體實施例中,不使用該等條件輸出反相器2〇6、21〇及該等 多工器207、211。 在另一技術中,半選擇字線及位元線可藉由取代該等單 一未選定偏壓線UXL及UYL而併入一個別反向源極選擇匯 流排來在該反向模式下提供。現在參考圖9,說明一字線 解碼器電路,其利用雙解碼源極選擇匯流排。用於該等字 線驅動器電路之該等PMOS電晶、體之一反向源極選擇匯流 排XSELP已取代圖4所示之未選定偏壓線UXL而併入。此 子線解碼器電路之剩餘部分如前述操作。 在該反向模式下,選定解碼輸出節點158係低態有效並 驅動至接地。該反向源極選擇匯流排XSELp之該等個別匯 123178.doc -24- 200826114Referring now to Figure 5, an exemplary bit line decoder circuit is shown including display of bias conditions suitable for a forward bias mode of operation (shown in Figure 1). A row of decoder circuits is displayed on the left side of the page, which displays two decoded outputs 2〇8, 212. Decoded output 208 corresponds to a selected decoded output and decoded output 212 corresponds to an unselected decoded output. A row of decoders 2 〇 2 can be implemented using any of a variety of well known techniques to produce a plurality of decoded outputs, such as outputs 205, 209, which are multiplexers 207, 211 and the inverters 2 〇 6, 210 To conditionally reverse. Unlike the column decoder, there is no inversion buffer to drive node 2〇5 after the gate, because the capacitive load on node 2〇8 is greater than the output for the column decoders. It is much lower. The row decoder 202 operates in this mode of operation, i.e., the upper supply voltage is coupled to the power supply node 203 and a grounded lower supply voltage is coupled to the power supply node 204. In this mode of operation, the row decoder is a low state / effect & decoder. The unselected decoded outputs (e. g., decoded output node 212) are driven to the highest of the two available voltage states, in this case VPP. The following will initially assume that only &lt;RTIgt; such decoding output node 208 (e.g., "low state") is selected. Each decoded output is coupled to - or a plurality of bit line driver circuits. For example, the decode output node 208 is coupled to a one-bit line driver circuit, which includes a PMOS transistor 221 and an NMOS transistor 222. The individual 汲 terminal sections of the transistors 221, 222 are simultaneously coupled to a single bit line, in this case the selected locating element line 106. Although a particular embodiment of the present invention encompasses a decoder other than a multi-head decoder, FIG. 5 depicts a second bit line driver circuit that is also coupled to the decode output node 208, which is associated with this particular decode output node 208. One or more remaining bit line driver circuits. The second word line driver circuit includes a PM 〇s transistor 223 & nm 〇s OLED 224. The output of the NMOS transistor is driven by a bit line 23 1 which represents a plurality of semi-selected positions it and a line. Comparing the word line decoder, the & class semi-selected positioning line can represent a selected positioning element line, which is maintained in an inactive state. The individual source terminals of the pM〇s transistors in each of the bit line driver circuits are coupled to a source select bus selB2 and a different bus bar. In this mode of operation, the source selection bus (3) is data dependent and can be further decoded based on the address information such that one or more such bus lines are suitable for this mode of operation. A bit line is biased in an active state, and for this mode of operation, the remaining bus bar is biased in an inactive state suitable for the bit line. In a particular embodiment, one or more of such source select bus bars may be active, but now assume that bus bar 217 is active and biased at vpp, and one or more remaining bus bars (It is indicated that the bus bar 2 is invalid and driven to the unselected positioning element line voltage VUB (shown as VT). Since the voltage (ground) on the decoded output node 208 is lower than the voltage of the bus bar lines 217, 218, Therefore, the ?? 8 transistors 221, 223 are both turned on, thereby driving the selected positioning element line 106 to VPP, and driving the semi-selected positioning elements 123178.doc -18-200826114 lines 23 1 to VT. The two conduction paths are indicated as open-ended arrow lines. The individual source terminals of the NMOS transistors in the various circuits of the bit line driver circuits are coupled to an unselected bias line UYL, also labeled as node 214. In this mode of operation, the UYL bias line passes the unselected positioning element line voltage VUB. Since the voltage (ground) on the decoded output node 208 is lower than the voltage of the UYL bias line, the second NMOS transistor 222, 224 is cut off. Decode output node 212 is coupled to a bit element A driver circuit comprising a PMOS transistor 225 and an NMOS transistor 226. The individual 汲 terminal sections of the transistors 225, 226 are simultaneously coupled to a bit line, in this case representing the unselected bit line 108. Also coupled to decoding A second bit line driver circuit of output node 212 represents one or more remaining bit line driver circuits associated with decode output node 212 and includes PMOS transistor 227 and NMOS transistor 228, the output of the transistors Driving an unselected positioning element line 233 ° As described above, the individual source terminals of the PMOS transistors in the respective circuits of the bit line driver circuits are coupled to one of the individual bus bars of a source selection bus line SELB. Since the voltage (VPP) on the decoded output node 212 is at or above the voltage of the bus bars 217, 218, the PMOSs 225, 227 are both turned off. Within the circuits of the bit line driver circuits. The individual source terminal of the NMOS transistor is coupled to the unselected bias line UYL node 214. Since the voltage on the decoded output node 212 is VPP, the two NMOS transistors 226, 228 are turned "on", thereby unselected positioning. The lines 108, 233 are driven to VUB (eg, VT). The two conductive paths are indicated by the 123178.doc -19-200826114 open-ended arrow line. Referring now to Figure 6, the bit line decoder circuit is shown, including The bias condition of the reverse dust operation mode (shown in Figure 2). The decoded output 208 of the row decoder circuit still corresponds to a selected decoded output, and the decoded output 2 12 corresponds to an unselected decoded output. The decoder 2〇2 operates in this mode of operation, with a supply voltage equal to the upper GND coupled to the power supply node 203 and a lower supply voltage _Vrr/2 coupled to the power supply node 2〇4. In this mode of operation, the row decoder is a &quot;high state valid&quot; decoder, and the active (selected) decoded output 208 is driven to two available voltage states by inverter 206 and multiplexer 207. The highest one, in this case, is GND (ground). The unselected decoded outputs (eg, decoded output node 212) are driven to the two available voltage states by inverter 210 and multiplexer 211. The lowest, in this case, is -VRR/2. In this mode of operation, for the exemplary embodiment, the individual busbars of the source select bus SELB are all driven to the same The bias condition (ground), and the "unselected" bias line UYL delivers a bias voltage equal to -VRR/2 (e.g., -5 volts). In this reverse mode of operation, the bias line UYL actually delivers an active state suitable for the bit line, rather than an invalid or unselected bias condition. Since the voltage (ground) on the decoded output node 208 is considerably higher than the voltage of the bias line uyL (below the NMOS threshold voltage), the NMOS transistors 222, 224 are both turned "on" to select the selected The bit line 106 is driven to -VRR/2 and also drives the half-selected bit line (shown here as the selected bit line 23A) to -VRR/2. The two conductive paths are indicated as open-ended arrow lines. 123178.doc -20- 200826114 In this mode of operation, the source select bus SELB is non-data dependent or not decoded (at least in a given block), and each such bus line is suitable for A bias of a bit line in an inactive state (eg, ground). Both of the PMOS transistors 221, 223 are turned off. Decode output node 212 is an unselected output and is driven to -VRR/2. The PMOS transistors 225, 227 are both turned "on" to drive the unselected clamp lines 108, 233 to ground. The two conduction paths are indicated as open-ended arrow lines. The two NMOS transistors 220, 228 are turned off. It should be noted that 'in this forward mode' the row decoder is active low and the bit το line is active high. However, in the reverse mode, the row decoder becomes active in the high state against its polarity, and the bit line itself is also inverted and becomes active in the low state. Conversely, in this forward mode, the column decoder is active high and the word lines are active low. However, in the reverse mode, the column decodes H to reverse its polarity and becomes active low, and the word lines themselves are reversed to be polar and become active. It should also be noted that the row decoder output level is offset above the average voltage between the forward mode (i.e., GND to VPP) and the reverse mode (i.e., _VRR/2 to GND). When viewed as a non-multi-head decoder (in Figures 3, 4, 5 and 6, only non-dashed array line driver circuits), the operation of the decoder circuit can be explained very simply. In the reverse mode, the word line decoder reverses its polarity and places a selected word line high (~5 V) while leaving all other selected word lines grounded. This reversal occurs on the bit field selection side, where the 敎-bit line is -5 V and all other bit lines are grounded. The end result is a 10 V reverse bias across the selected memory cell and a zero reverse bias across the other cells. 123178.doc -21 - 200826114 These transistors in the sub-line and bit line drive II circuits only have to withstand 5 V, or half the maximum voltage, not the entire voltage. When considering the use of a multi-headed stone device (in Figures 3, 4, 5 and 6, including the virtual array line driver circuit), it should be noted that the circuit thus far uses the -decoding source in the forward direction. A bus is selected that allows selection of the single group of array lines (but simultaneously drives the remaining semi-selected array lines to an unselected (four) condition). However, in the reverse mode, the selected decoded output from the column and row decoders couples the array lines to a single unselected bias line (e.g., UXL and UYL). Semi-selected array lines are not known under this inverse weight using the m-line. Thus, the above described circuits and techniques are useful in configuring an array of line blocks (e.g., block erase,) to be selected in the reverse mode. As can be seen in Figures 4 and 6, a selected wordline block and a selected locating meta-line block are simultaneously selected in the reverse mode without any independently configurable semi-selected array lines. This type of block operation avoids the need for any semi-selected lines. The decoding implication may be very similar to the word line configuration having a multi-layer word line segment for a three-dimensional memory array, as described in U.S. Patent No. 6,879,5,5, issued to </RTI> &lt;heUerlein, which is incorporated herein by reference. Reveal the inner valley king.并入 is incorporated herein by reference. Whether this block operation can be configured (or how large a block can be configured) depends mainly on the number of cell reset currents, the number of cells that conduct such reset currents, and the word line driver circuit and bit lines. Whether the PM〇s and NMOS transistors in the driver circuit can support such currents at an acceptable voltage drop. A semi-selected array line can be provided in this reverse mode by using other techniques (except in the forward mode). In a single such technique, 123178.doc -22-200826114, the column and row decoders may be powered by an overvoltage such that the decoded output nodes are above the PMOS source voltage and below the NM0S source voltage. across. By doing so, the selected word line can be driven to the south by +NVR/2 voltage through the NM〇s transistor, and the selected positioning line can be driven down to the -VRR/2 voltage through the PM〇s transistor. This utilizes the same transistor as during the forward mode to drive the selected word line and bit line. Such techniques are illustrated in Figures 7 and 8. Referring initially to Figure 7, a word line decoding is illustrated as a circuit that utilizes an overdrive decode output to drive the array line drivers with their sources maintained under the bias conditions described above. In this column decoder circuit, column decoder 152 is powered by an 8 volt upper supply voltage and a negative i volt lower supply voltage. The polarities of the decoded output nodes 158, 162 are inversely reversed with respect to the polarity shown in Figure 4, and are now a high effective decoder that provides a selected output 158 at +8 volts and provides one at _ volts. Decoded output 162 is not selected. The source select bus xsel keeps a decoding busbar unchanged. The (or more) of its individual flow lines are selected and driven to +5 volts, and the unselected lines are driven to ground. The nm〇s transistor 172 is turned "on" and conducts the selected word line 1〇2 to the associated bus line voltage (+5 volts). The NM〇s transistor m is also turned "on" and conducts the (equal) half-selected word line 181 to ground. When the decoded output node is not selected, the PMOS transistors 175, 177 are simultaneously turned on, and the unselected W lines πμ, 183 are conducted to ground. In some specific implementations using this technique, the conditions (4) and the like are not output to the inverters 156, 160 and the multiplexers 157, 161 (shown here as "dotted lines"). Referring now to Figure 8, a description is made. A bit line decoder circuit that also drives the array line drivers using a 123178.doc -23 - 200826114 overdrive decoding output. In this row decoder circuit, the row decoder 202 is powered by a +1 volt power supply. The voltage is supplied with a negative supply voltage of minus 8 volts. The polarities of the decoded output nodes 2 〇 8, 212 are inversely opposite to the polarity shown in Figure 6, and are now a low effective decoder at _8 volts. A selected output 208 is provided and an unselected decoded output 212 is provided at +1 volt. One (or more) of the individual 8 £1^6 bus bars 217 are selected and driven to _5 volts. The unselected 16 bus bars 218 are driven to ground. The PM〇s transistor 221 is turned "on" and conducts the nominal select bit line 1 〇 6 to the associated SELB bus line voltage (_5 volts). The PMOS transistor 223 is also turned "on" and conducts the (equal) semi-selected word line 23 1 to When the unselected decode output node 212 is at +1 volt, the NMOS transistors 226, 228 are simultaneously turned on and the unselected locating elements 108, 233 are conducted to ground. In some embodiments, the conditional output inverters 2〇6, 21〇 and the multiplexers 207, 211 are not used. In another technique, the semi-selected word lines and bit lines can be replaced by The single unselected bias lines UXL and UYL are combined into a different reverse source select bus to be provided in the reverse mode. Referring now to Figure 9, a word line decoder circuit is utilized which utilizes a dual decoding source. A pole selection bus. One of the PMOS transistors for the word line driver circuits, the reverse source select bus XSELP has been incorporated in place of the unselected bias line UXL shown in FIG. The remainder of the line decoder circuit operates as previously described. In this reverse mode, the selected decode output node 158 is active low and driven to ground. The individual sources of the reverse source select bus XSELp are 123178.doc - 24-200826114

流排線之一選定者係偏壓至一適用於一字線之反向操作模 式之有效偏壓條件。在此情況下,該XSELP匯流排之選定 匯流排線243係驅動至VRR/2,而該XSELP匯流排之未選定 偏壓線244係驅動至一適用於一字線之此操作模式之無效 偏壓條件,在此情況下係驅動至接地。PM〇s電晶體171係 藉由麵合至其閘極之低壓而接通,並將選定字線1〇2驅動 至VRR/2電位。然而,在該半選定字線驅動器電路内的 PMOS電晶體173保持截止,因為在其閘極上的電壓相對於 其源極不夠低,由於二者均接地。 由於NMOS電晶體174係也截止,故在該半選定字線驅動 器電路内的任一電晶體均不接通。因此,該等半選定字線 在接地電位或附近浮動。如同在範例性電路之情況下,在 NMOS下拉電晶體174大於pM〇s上拉電晶體173時發生此 情況。更大電晶體比更小電晶體完全具有一至其基板更大 洩漏數量。因此,由於電晶體174具有一捆綁至接地之基 板,故接地洩漏電流支配基板洩漏電流至由pM〇s電晶體 173所產生之VRR/2,且此淨電流傾向於將該等未選定字 線⑻維持在接地電位或附近。與未選定解碼輸出節點162 相關聯之該等字線㈣器電路如前述操作,該等nm〇s電 晶體176、178係接通以將該等未選定字線104、183傳導至 接地。 在一替代性具體實施例中 之低位準可藉由使用一等於 作列解碼器152、反相器i56 ’该專解碼輸出節點1 5 8、1 6 2 -ντρ(或更低)之低電源154操 、160及多工器157、ι61來驅 123178.doc -25- 200826114 動至低於接地(例如至一在接地以下PMOS臨界電壓或以下 之電壓’即-VTP)。由此,PM0S上拉電晶體173係接通至 有效驅動該(等)半選定字線1 8 1至接地。 一類似情形發生於併入雙資料相依源極選擇匯流排之一 行解碼器電路中。現在參考圖丨〇,說明一位元線解碼器電 路’其利用雙解碼(在此情況下資料相依)源極選擇匯流 排。用於該等位元線驅動器電路之該等NMOS電晶體之一 反向源極選擇匯流排SELN已取代圖6所示之未選定偏壓線 UYL而併入。此位元線解碼器電路之剩餘部分如前述操 作。 在該反向模式下,選定解碼輸出節點208係高態有效並 驅動至接地。該反向源極選擇匯流排SELN之該等個別匯 流排線之一選定者係偏壓至一適用於一位元線之反向操作 模式之有效偏壓條件。在此情況下,該SELN匯流排之選 定匯流排線247係驅動至-VRR/2,而該SELN匯流排之未選 定偏壓線248係驅動至一適用於此操作模式之位元線之無 效偏壓條件,在此情況下係驅動至接地。NMOS電晶體222 係藉由耦合至其閘極之高電壓而接通,並將選定字線1 〇6 驅動至-VRR/2電位。然而,在該半選定字線驅動器電路内 的NMOS電晶體224保持截止,因為在其閘極上的電壓相對 於其源極不夠高,由於二者均接地。 由於PMOS電晶體223係也截止,故在該半選定位元線驅 動器電路内的任一電晶體均不接通。因此,該等半選定位 儿線在接地電位或附近浮動。如同在此範例性電路之情況 123178.doc -26- 200826114 下,在PMOS上拉電晶體223大於NMOS下拉電晶體224時 發生此情況。更大電晶體比更小電晶體完全具有一至其基 板之更大洩漏數量。因此,由於更大電晶體223具有一捆 綁至接地之基板,故接地漏電流支配基板洩漏電流至由 NMOS電晶體224所產生之_vrr/2,且此淨電流傾向於將 該等半選定字線23 1維持在接地電位或附近。與未選定解 碼輸出節點212相關聯之該等位元線驅動器電路如前述操 作,該等PMOS電晶體225、227係接通以將該等未選定位 元線108、233傳導至接地。 對於該等解碼器電路二者,在該正向操作下的操作實質 上如圖4及6所示而執行。考量該列解碼器情況,在該正向 模式下,解碼該源極選擇匯流排,並將所有未選定字線驅 動至該未選定偏壓線UXL。在使用該雙解碼列解碼器之正 向模式下,不解碼該反向源極選擇匯流排,並將所有其個 別匯流排線驅動至與該UXL匯流排線相同之電壓。因而, 該等字線驅動器電路相對於圖4不變地操作。確實,一單 一偏壓線UXL已由複數個”偏壓線”取代,各偏壓線係驅動 至與前者UXL·線相㈣電壓,且各未選定字線係驅動 至該偏壓線。 在該行解碼器情況下’在該正向模式下解碼該源極選擇 匯流排SELB ’並將所有未選定位元線驅動至該未選定偏 I線UYL。在使㈣雙解碼行解碼器之正向模式下,不解 碼該反向源極選擇㈣排,並將所有其個龍流排線驅動 至與該UYL匯流排線相同的電屢。因而,該等位元線驅動 123178.doc •27- 200826114One of the streamlines is selected to be biased to an effective bias condition suitable for the reverse mode of operation of a word line. In this case, the selected bus bar 243 of the XSELP bus is driven to VRR/2, and the unselected bias line 244 of the XSELP bus is driven to an invalid bias for this mode of operation for a word line. The pressure condition, in this case, is driven to ground. The PM〇s transistor 171 is turned on by the low voltage that is turned to its gate, and drives the selected word line 1〇2 to the VRR/2 potential. However, the PMOS transistor 173 within the half-selected wordline driver circuit remains off because the voltage on its gate is not sufficiently low relative to its source, since both are grounded. Since the NMOS transistor 174 is also turned off, none of the transistors in the half-selected word line driver circuit are turned "on". Therefore, the semi-selected word lines float at or near the ground potential. As in the case of the exemplary circuit, this occurs when the NMOS pull-down transistor 174 is larger than the pM 〇s pull-up transistor 173. Larger transistors have a larger number of leaks than their smaller ones. Therefore, since the transistor 174 has a substrate tied to the ground, the ground leakage current dominates the substrate leakage current to the VRR/2 generated by the pM〇s transistor 173, and this net current tends to place the unselected word lines. (8) Maintain at or near the ground potential. The word line (four) circuits associated with the unselected decode output node 162 operate as previously described, and the nm 〇s transistors 176, 178 are turned "on" to conduct the unselected word lines 104, 183 to ground. The low level in an alternative embodiment can be achieved by using a low power supply equal to the column decoder 152, the inverter i56 'the dedicated decoding output node 1 5 8 , 1 6 2 -ντρ (or lower). 154, 160, and multiplexers 157, ι 61 drive 123178.doc -25- 200826114 move to below ground (for example, to a voltage below the ground PMOS threshold voltage or below 'that is -VTP). Thus, the PMOS pull-up transistor 173 is turned "on" to effectively drive the (equal) half-selected word line 181 to ground. A similar situation occurs in a row decoder circuit that incorporates a dual data dependent source select bus. Referring now to Figure 丨〇, a bit line decoder circuit is illustrated which utilizes dual decoding (in this case data dependent) source selection bus. One of the NMOS transistors for the bit line driver circuit, the reverse source select bus SELN, has been incorporated in place of the unselected bias line UYL shown in FIG. The remainder of this bit line decoder circuit operates as previously described. In this reverse mode, the selected decode output node 208 is active high and driven to ground. One of the individual bus bars of the reverse source select bus SELN is biased to an effective bias condition suitable for the reverse mode of operation of the one bit line. In this case, the selected bus bar 247 of the SELN bus is driven to -VRR/2, and the unselected bias line 248 of the SELN bus is driven to a bit line suitable for this mode of operation. The bias condition, in this case, is driven to ground. NMOS transistor 222 is turned "on" by a high voltage coupled to its gate and drives selected word line 1 〇6 to the -VRR/2 potential. However, the NMOS transistor 224 in the half-selected word line driver circuit remains off because the voltage across its gate is not sufficiently high relative to its source, since both are grounded. Since the PMOS transistor 223 is also turned off, any of the transistors in the half-selected locator driver circuit are not turned "on". Therefore, the semi-selected positioning lines float at or near the ground potential. As in the case of this exemplary circuit, 123178.doc -26-200826114, this occurs when the PMOS pull-up transistor 223 is larger than the NMOS pull-down transistor 224. Larger crystals have a much larger number of leaks than their smaller ones than smaller ones. Therefore, since the larger transistor 223 has a substrate tied to the ground, the ground leakage current dominates the substrate leakage current to _vrr/2 generated by the NMOS transistor 224, and the net current tends to be such a semi-selected word. Line 23 1 is maintained at or near the ground potential. The bit line driver circuits associated with the unselected decode output node 212 operate as described above, and the PMOS transistors 225, 227 are turned "on" to conduct the unselected bit lines 108, 233 to ground. For both decoder circuits, the operation under this forward operation is substantially performed as shown in Figures 4 and 6. Consider the column decoder case, in which the source select bus is decoded and all unselected word lines are driven to the unselected bias line UXL. In the forward mode using the dual decode column decoder, the reverse source select bus is not decoded and all of its individual bus bars are driven to the same voltage as the UXL bus. Thus, the word line driver circuits operate unchanged with respect to FIG. Indeed, a single bias line UXL has been replaced by a plurality of "bias lines" that are driven to the former UXL. line phase (four) voltage and each unselected word line is driven to the bias line. In the case of the row decoder, the source select bus SELB&apos; is decoded in the forward mode and all unselected bit lines are driven to the unselected off-line UYL. In the forward mode of the (four) double decoding row decoder, the reverse source selection (four) row is not decoded, and all of its respective bus lines are driven to the same electrical frequency as the UYL bus line. Thus, the bit line driver 123178.doc •27- 200826114

器電路相對於圖6不變地操作。確實,一單一偏壓線UYLThe circuit operates unchanged with respect to Figure 6. Indeed, a single bias line UYL

已由複數個”偏壓線”取代,各偏壓線係驅動至與前者UYL 偏壓線相同之電壓’且各未選定位元線係驅動至該偏壓 線。 至此所述的該等解碼器電路係用於實施其中記憶體單元 包括一可逆電阻器加上一二極體之記憶體陣列。此類記憶 體單元可使用橫跨單元施加之一反向偏壓來重置,且用於 半選定字線及位元線允許將個別字線及位元線放置於一重 置偏壓條件下,從而提供重置個別記憶體單元而不須重置 一整個區塊之能力。 如圖7及8所述之技術具有僅一單一解碼源極選擇匯流排 之優點,但由於該等列及行解碼器係由過電壓來供電,故 用於此類解碼器電路之電壓要求更高。圖9及1〇所述之技 術在-額外解碼(及/或資料相依)反向源極選擇匯流排及併 入使用二解碼源極選擇匯流排之陣列線驅動器之可能増加 面積的代價下’藉由不利用過電壓向該等二解碼器電路供 電來減小該等電壓要求。該位元線選擇電路多達匯流排線 的兩倍’且可能佈線受限。該等字線選擇電路還可能略微 更大且佈線受限(即該等字線驅動器電路包括六個額外解 碼線詩—六頭解碼器,且該pM㈣置略微大於更早期 的電路)。耗如此,但任—技術可能比用於特定具體實 施例之其他技術更有用。 上面在一程式化條件之昔县 ^干您才不下說明該正向模式,其中施 加至該選定位元線之電壓# vp _ ^ , 电座係VPP。该正向模式還應用於一 123178.doc -28- 200826114 讀取模式,其中選定位元線係驅動至一讀取電壓Vrd,且 選定字線再次驅動至接地。此類讀取電壓可以係一比該程 式化電壓VPP低得多的電壓,且該未選定字線偏壓電壓 VUX及該未選定位元線偏壓電壓VUB因此在其用於該程式 化模式之值上減小。 特定記憶體單元可使用一正向偏壓模式來加以”程式化,,, 並使用該反向模式來抹除區塊。其他單元可使用一最初正 向偏壓程式化技術來預調節(例如在製造期間),但接著使 用該反向模式來加以”程式化”,並使用該正向模式來加以 &quot;抹除π。為了避免與可程式化技術中的歷史用法混淆,並 為了全面瞭解搭配至此所述之該等解碼器電路使用所構思 之不同記憶體技術,三個不同操作模式係用於說明:讀 取、設定及重置。在該讀取模式下,橫跨一選定記憶體單 元施加一讀取電壓VRD。在該設定模式下,橫跨一選定記 憶體單元施加一設定電壓VPP。在至此所述之範例性具體 實施例中,該讀取電壓VRD及該設定電壓Vpp二者均係正 電壓’且此模式係使用正向解碼器操作模式來實施。在該 重置模式下’松跨一選定記憶體單元施加一重置電壓 VRR。在至此所述之範例性具體實施例中,該重置電壓 VRR係作為一反向偏壓電壓來施加,並使用該反向解碼器 操作模式來實施。 上述重置模式使用一分割電壓技術來限制用於該等解碼 器電路之該等電壓要求,並將一選定位元線驅動至一負電 壓(即使用一三重井半導體結構)。或者,該重置模式可使 123178.doc -29- 200826114 用完全非負電壓來實施。在此情況下,該重置電壓Vrr係 傳遞至該選定字線,且接地係傳遞至該選定位元線。該等 VUX及VUB電壓較佳的係設定至大約VRR/2。 許多類型的記憶體單元(下述)能夠使用該重置模式來加 以程式化。在該些記憶體單元技術之特定技術中,在各記 憶體單元内的一反熔絲最初在正向方向上跳變。接著在反 向偏壓方向上”調諧,,各記憶體單元之電阻以完成程式化。 此將對於一一次可程式化單元亦如此情況。對於可再寫單 元,使用該正向方向來抹除單元,其可在一各種大小之區 塊内執行,接著使用該反向模式來加以程式化。 該反向偏壓係用於重置該選定記憶體單元。該程式化電 Μ係由一二極體崩潰來供應。此外,可仔細控制與此程式 化相關聯之該等偏壓條件,包括控制該選定字線及/或位 元線之電壓斜坡。有用程式化技術之額外洞察可見諸於下 面所參考之美國專利第6,952,03 0號。如下面所參考之〇23_ 0049及023-0055申請案所述,以及如下面所參考之μα_ 163-1申請案中更詳細所述,多個程式化操作可用於程式 化各種電阻狀態。傾斜程式化脈衝之用途係說明於下面所 參考之SAND-OimuSO及SAND-01114US1申請案内,且用 於微調多個單元之電阻之技術係說明於下面所參考之 SAND-01117US0及 SAND-01117US1 申請案内。 如上述(特別在雙解碼源極選擇線之背景下)使用重置程 式化來程式化一併入一可微調電阻元件之被動元件記憶體 單元在提供較大彈性以允許一更大陣列區塊大小時特別有 123178.doc •30- 200826114 用。即便在一選定陣列區塊内(如全部上述曾假定),在該 重置模式下仏跨咸荨未選定記憶體單元不存在任何偏壓, 因此沒有浪費的功率消耗。透過一單元之反向電流(Irev) 與區塊大小無關。因此可選擇許多區塊以增加寫入帶寬。 此外,桉跨各半選定記憶體單元之電壓僅係該程式化電壓 之一半’故安全地用於該些單元。Already replaced by a plurality of "bias lines", each bias line is driven to the same voltage as the former UYL bias line and each unselected bit line is driven to the bias line. The decoder circuits described so far are used to implement a memory array in which the memory unit includes a reversible resistor plus a diode. Such a memory cell can be reset using a reverse bias applied across the cell, and for semi-selected word lines and bit lines allows individual word lines and bit lines to be placed under a reset bias condition To provide the ability to reset individual memory cells without having to reset an entire block. The techniques described in Figures 7 and 8 have the advantage of having only a single decoded source select bus, but since the column and row decoders are powered by overvoltage, the voltage requirements for such decoder circuits are even greater. high. The techniques described in Figures 9 and 1B are at the expense of an additional decoding (and/or data dependent) reverse source select bus and a possible add-on area incorporating an array line driver using two decoded source select buses. These voltage requirements are reduced by not supplying overvoltages to the two decoder circuits. This bit line selection circuit is up to twice as large as the bus bar and may be limited in wiring. The word line select circuits may also be slightly larger and have limited wiring (i.e., the word line driver circuits include six additional decoding line poems - six header decoders, and the pM (four) is slightly larger than the earlier circuits). This is the case, but the technology may be more useful than other techniques for a particular embodiment. Above the above-mentioned stylized conditions, you can't explain the forward mode, where the voltage applied to the selected location line is #vp _ ^ and the battery is VPP. The forward mode is also applied to a read mode of 123178.doc -28-200826114, wherein the selected bit line is driven to a read voltage Vrd and the selected word line is driven again to ground. Such a read voltage may be a voltage that is much lower than the stylized voltage VPP, and the unselected word line bias voltage VUX and the unselected bit line bias voltage VUB are thus used in the stylized mode The value is reduced. A particular memory cell can be "programmed" using a forward bias mode, and the reverse mode is used to erase the block. Other cells can be pre-conditioned using an initial forward bias programming technique (eg, During manufacturing, but then use the reverse mode to "stylize" and use the forward mode to &quot;wipe π. To avoid confusion with historical usage in stylized techniques, and to fully understand The decoder circuits described so far use the different memory technologies conceived, and three different modes of operation are used to illustrate: read, set, and reset. In this read mode, across a selected memory The cell applies a read voltage VRD. In the set mode, a set voltage VPP is applied across a selected memory cell. In the exemplary embodiment described so far, the read voltage VRD and the set voltage Vpp are Both are positive voltage' and this mode is implemented using a forward decoder mode of operation. In this reset mode, a reset voltage VRR is applied across a selected memory cell. In the exemplary embodiment, the reset voltage VRR is applied as a reverse bias voltage and implemented using the reverse decoder mode of operation. The reset mode is limited using a split voltage technique. The voltage requirements of the decoder circuits drive a selected positioning line to a negative voltage (ie, using a triple well semiconductor structure). Alternatively, the reset mode can be used for 123178.doc -29-200826114 The reset voltage Vrr is transmitted to the selected word line, and the ground is transmitted to the selected bit line. The VUX and VUB voltages are preferably set to approximately VRR/. 2. Many types of memory cells (described below) can be programmed using this reset mode. In the particular technique of memory cell technology, an antifuse in each memory cell is initially in the forward direction. The direction jumps. Then, in the reverse bias direction, "tuning, the resistance of each memory cell to complete the stylization. This will be the case for a once programmable unit. For rewritable cells, the forward direction is used to erase the cells, which can be executed within blocks of various sizes, and then programmed using the reverse mode. The reverse bias is used to reset the selected memory unit. The stylized motor is supplied by a diode collapse. In addition, the bias conditions associated with the programming can be carefully controlled, including controlling the voltage ramp of the selected word line and/or bit line. Additional insights into useful stylization techniques can be found in U.S. Patent No. 6,952,030, which is incorporated herein by reference. A number of stylized operations can be used to program various resistance states, as described in more detail in the applications of the pp. 23- 0049 and 023-0055, the disclosure of which is incorporated herein by reference. The use of tilted stylized pulses is described in the SAND-OimuSO and SAND-01114US1 applications referenced below, and the techniques for fine-tuning the resistance of a plurality of cells are described in the SAND-01117US0 and SAND-01117US1 applications referenced below. . As described above (especially in the context of dual decoded source select lines), the use of reset programming to program a passive component memory cell incorporating a trimmable resistive element provides greater flexibility to allow for a larger array of blocks. Large hours are especially for 123178.doc •30- 200826114. Even within a selected array block (as all have been assumed above), there is no bias in the reset memory mode for the unselected memory cells, so there is no wasted power consumption. The reverse current (Irev) through a cell is independent of the block size. Therefore many blocks can be selected to increase the write bandwidth. In addition, the voltage across the selected memory cells is only half of the programmed voltage' and is safely used for the cells.

應注意,在上述中,該重置模式說明選定及半選定字線 及位元線。例如在列選擇之背景下,一給定位址可實際上 不k疋此類半選定字線,且此術語係該多頭字線驅動器 結構之一人為產物。然而,在該等位元線之背景下,此類 半選疋位70線可能實際上只要與行位址有關便可選定,但 可此偏壓至一用於該等位元線之無效狀態而非有效狀態, 口為用於忒位元線之特定資料不需要”程式化,,單元,或因 為β位7L線正在’’等待„被程式化。此情況在同時程式化少 於位元線解碼器頭數目時發生。然而應注意,程式化帶寬 顧慮提出組態一記憶體陣列以盡可能多地同時程式化位元 Λ 許該(等)敎位元賴得—負電壓,而該 (等)選疋子線獲得一正電壓。在重置程式化(即反向模式) 下,用於全部未選定陣列線(未位元線及字線)之參考位準 係接地1而快速解嗎及選擇字線及位元線二者。再次: 考該等半敎字線及位元線係㈣在接地(由於至今等: 驅動器'晶體之最大者之井電位㈣漏電流)之說二: 等Z it體早疋之電㈣f在此類半選定陣列線與該等未選 123178.doc 200826114 定陣列線之間提供-額外洩漏電流,等陣列線係主動維 持在該未選定偏壓位準下。此進一步促進該等未選定陣列 線保持在該未選定偏壓電位或附近浮動。 涵蓋二維記憶體陣列,但咸信該等解碼器配置對於一具 有多個記憶體平面之三維記憶體陣列特別有用。在特定較 佳具體實施例中,該記憶體陣列係組態使得各字線在多個 記憶體平面之各記憶體平面上包含字線片段,如下所述。 圖11係一範例性記憶體陣列3〇〇之一方塊圖。雙列解碼 器3 02 304產生列選擇線用於該陣列,各列選擇線橫跨陣 列3 00,如本文下面所述。在此具體實施例中,該等字線 驅動器電路(未顯示)空間分佈於該記憶體陣列下面並藉助 在個別記憶陣列區塊(標註3〇6、3〇8的兩個區塊)之交替側 上的垂直連接(其中一個係標註3 1〇)來連接至該等字線。所 示記憶體陣列包括兩個記憶體,,條”3 18、32〇,並進一步包 括四個行解碼器及分別在該陣列之頂部、中上部、中下部 及底部的位元線電路區塊312、314、315、316。如本文所 述,還可併入額外條,且各條可包括一或多個記憶體機 架。在各區塊内的該等位元線還2:丨交錯以鬆弛行相關電 路之間距要求。作為一範例,位元線322與上部行電路區 塊3 12相關聯(即由其驅動並感應),而位元線324係與底部 行電路區塊3 14相關聯。 在範例性具體實施例中,記憶體陣列3〇〇係在四個記憶 體平面之各記憶體平面上形成之被動元件記憶體單元之一 二維記憶體陣列。此類記憶體單元較佳的係併入一可微調 123178.doc -32- 200826114 電阻器元件(如本文所述),並還可包括一反熔絲。各邏輯 字線係連接至在四個字線層之各字線層上的一字線片段 (各與一個別記憶體平面相關聯)。 記憶體陣列300之各條係分成大量區塊,例如區塊308。 在本文所述之特定範例性具體實施例中,各記憶體機架包 括16個卩車列區塊,但可實施其他數目的區塊。在所示範例 性具體實施例中,各區塊在用於個別四個記憶體平面之四 個位元線層之各位元線層上包括288個位元線,因而總計 每區塊1,1 52個位元線。該些位元線係2:1交錯,使得在一 陣列區塊之頂部及底部的該等行解碼器及資料1/〇電路之 各電路介接576個位元線。還涵蓋此類位元線及陣列區塊 之其他數目及配置,包括更高數目。 在一選定記憶體陣列區塊内,該些源極選擇匯流排線 XSELN(或反向源極選擇匯流排XSELp)之一係由一列偏壓 電路來解碼並驅動至一有效偏壓條件,而剩餘匯流排線 (還稱為”偏壓線”)係驅動至一無效條件(即一適用於一未選 定字線之電壓)。因此,一單一選定RSEI^.(即列選擇線, 其對應於圖3内的解碼輸出節點158)驅動該選定記憶體區 塊内的一字線為低態,並將該選定區塊内的其他Ν」字線 驅動至一未選定偏壓位準。在其他非選定記憶體區塊内, 3等源極及反向源極選擇匯流排之任何個別匯流排線均不 驅動為有效,故該有效RSEL線不選擇任何字線。或者, 可使在未選定陣列區塊内的該等源極及反向源極選擇匯流 排浮動,特別係在該正向模式下。 123178.doc -33 - 200826114 各列選擇線橫跨整個記憶體條内的所有記憶體區塊,並 於該條之各對區塊(以及兩個以上,各分別位於該 二-及最後區塊&quot;外部之間&quot;的一個別四頭字線驅動 裔口亥等RSEL線還可稱為”全域列線”,且還可對應於本文 所參考之該等列解碼器輸出節點。範例性電路、操作、偏 昼條件、浮動條件、操作模式(包括讀取及程式化模式)及 類似等之額外細節進—步說明於前述美國專利第◎乃,奶 號丄並另外說明於授予christopher L Petti等人之美國專利 案第7,〇54,219號,標題為&quot;用於緊密間距記憶體陣列線之 電晶體佈局組態&quot;’其全部揭示内容以引用形式併入本 文,並進一步說明於Roy Ε· Scheuerlein等人在2〇〇5年6月7 日申請的美國中請案第11/146,952號,標題為,·詩非二進 制記憶體線驅動器群組之解碼電路&quot;,作為美國專利申請 公告案第2006-022^02號於2006年10月5日頒佈,其全= 揭示内容以引用形式併入本文。 為了加快一全域列線之選擇時間,該些RSEI^^係藉由 二階層列選擇解碼器3 〇 2、3 〇 4 (還稱為,,全域類解碼器 302、304&quot;),各分別在陣列條左右側而位於陣列外部,在 其兩端來加以驅動。藉由使用一階層解碼器結構,減小全 域列解碼器302之大小,從而改良陣列效率。此外,可方 便地提供一反向解碼模式以獲得改良測試能力,如進—步 說明於2006年7月6日作為美國抓了申請公告案第2〇〇6_ (H45193號頒佈,於2〇04年12月3〇日申請的美國申請案第 11/026’493號’ Kenneth K· So等人之&quot;雙模式解碼器電路、 123178.doc -34- 200826114 併入其之積體電路記憶體陣列及相關操作方法&quot;中,立全It should be noted that in the above, the reset mode illustrates selected and semi-selected word lines and bit lines. For example, in the context of column selection, a given address may not actually be such a semi-selected word line, and this term is an artifact of the multi-headline driver structure. However, in the context of the bit line, such a half-selected 70 line may actually be selected as long as it relates to the row address, but may be biased to an inactive state for the bit line. Instead of the valid state, the port does not need to be "programmed," for the specific data used for the bit line, or because the beta bit 7L line is 'waiting' to be stylized. This happens when the number of bit line decoder headers is less than programmed at the same time. It should be noted, however, that stylized bandwidth concerns suggest configuring a memory array to simultaneously program as many bits as possible to allow the (equivalent) bit to depend on the negative voltage, and the (equal) selected sub-wires are obtained. A positive voltage. In reset stylization (ie, reverse mode), the reference bits for all unselected array lines (no bit lines and word lines) are grounded 1 and quickly solved and both word lines and bit lines are selected. . Again: test the semi-敎 word line and bit line system (4) in the grounding (because of the current: the driver's largest crystal well potential (four) leakage current) said two: the Z it body early electricity (four) f here An additional leakage current is provided between the semi-selected array lines and the unselected array lines, and the array lines are actively maintained at the unselected bias level. This further facilitates that the unselected array lines remain floating at or near the unselected bias potential. Two-dimensional memory arrays are covered, but it is believed that these decoder configurations are particularly useful for a three-dimensional memory array with multiple memory planes. In a particularly preferred embodiment, the memory array is configured such that each word line includes word line segments on respective memory planes of a plurality of memory planes, as described below. Figure 11 is a block diagram of an exemplary memory array. Dual column decoder 3 02 304 generates column select lines for the array, with column select lines spanning array 3 00, as described herein below. In this embodiment, the word line driver circuits (not shown) are spatially distributed under the memory array and alternate by means of individual memory array blocks (two blocks labeled 3〇6, 3〇8). Vertical connections on the sides (one of which is labeled 3 1〇) are connected to the word lines. The illustrated memory array includes two memories, strips "3 18, 32", and further includes four row decoders and bit line circuit blocks at the top, middle, bottom, bottom, and bottom of the array, respectively. 312, 314, 315, 316. Additional strips may also be incorporated as described herein, and each strip may include one or more memory racks. The bit lines within each block are also 2: interlaced As an example, the bit line 322 is associated with (ie, driven and sensed by) the upper row circuit block 312, and the bit line 324 is tied to the bottom row circuit block 3 14 . In an exemplary embodiment, the memory array 3 is a two-dimensional memory array of passive element memory cells formed on the memory planes of the four memory planes. Preferably, a resistor element (as described herein) can be fine-tuned 123178.doc -32-200826114, and can also include an anti-fuse. Each logic word line is connected to each of the four word line layers. a line segment on the word line layer (each with a different memory plane) Correlation) Each strip of memory array 300 is divided into a plurality of blocks, such as block 308. In the particular exemplary embodiment described herein, each memory rack includes 16 brake train blocks, but Implementing other numbers of blocks. In the exemplary embodiment shown, each block includes 288 bit lines on each of the bit line layers for the four bit line layers of the individual four memory planes, thus A total of 1,1 52 bit lines per block. These bit lines are 2:1 interleaved so that the row decoders and data 1/〇 circuits of the top and bottom of an array block are interposed. 576 bit lines are also included. Other numbers and configurations of such bit lines and array blocks, including higher numbers, are also contemplated. Within a selected memory array block, the source selects bus lines XSELN (or One of the reverse source select busses XSELp) is decoded and driven to a valid bias condition by a column of bias circuits, while the remaining bus bars (also referred to as "bias lines") are driven to an inactive condition. (ie one applies to the voltage of an unselected word line). Therefore, a single selected R SEI^. (ie, the column select line, which corresponds to the decode output node 158 in FIG. 3) drives a word line in the selected memory block to a low state, and the other Ν" word lines within the selected block. Drive to an unselected bias level. In other non-selected memory blocks, any individual bus lines of the 3 source and reverse source select bus bars are not driven to be active, so no valid word lines are selected for the active RSEL line. Alternatively, the source and reverse source select busses within the unselected array block may be floated, particularly in the forward mode. 123178.doc -33 - 200826114 Each column selection line spans all memory blocks in the entire memory strip, and each pair of blocks in the strip (and more than two, each located in the second and last block) The RSEL line, such as the “external column line”, may also be referred to as the “global column line” and may also correspond to the column decoder output nodes referenced herein. Additional details of operations, eccentric conditions, floating conditions, operating modes (including reading and stylized modes), and the like are described in the aforementioned U.S. Patent No. ◎, Milk No. 丄 and additionally stated in christopher L Petti U.S. Patent No. 7, pp. 54,219, entitled &quot;Crystal Layout Configuration for Tightly Spaced Memory Array Lines&quot;, the entire disclosure of which is hereby incorporated by reference herein in Ε· Scheuerlein et al., U.S. Patent Application No. 11/146,952, filed on June 7, 2005, entitled "Decoding Circuit of Poetry Non-Binary Memory Line Driver Group" as a US Patent Application public Case No. 2006-022^02 was promulgated on October 5, 2006, and its full disclosure is incorporated herein by reference. To speed up the selection of a global column line, the RSEI^^ is based on two hierarchical columns. Select decoders 3 〇 2, 3 〇 4 (also referred to as global domain type decoders 302, 304 &quot;), each located outside the array on the left and right sides of the array strip, and driven at both ends thereof. The hierarchical decoder structure reduces the size of the global column decoder 302 to improve array efficiency. In addition, a reverse decoding mode can be conveniently provided to obtain improved test capabilities, such as the step-by-step instructions on July 6, 2006. The United States arrested the application for the publication of the second paragraph 6_6 (promulgated by H45193, US application No. 11/026'493 filed on December 3, 2004, 'Kenneth K· So et al.' The decoder circuit, 123178.doc -34- 200826114 is incorporated into its integrated circuit memory array and related operating methods.

部㈣内容以引用方式併入本文。用於此類階層式解碼器 之乾例性電路可f諾…M Λ1 了見渚於美國專利申請公告案第2006- 0146639 Α1 號,Luca G Fas〇li犛人夕,,m • 〇h等人之用於使用多層級多 頭解碼器之密集記憶體陣列之階層解碼之裝置及方法&quot;, 其全部揭示内容係以引用形式併入本文。 在本文所述之特定材料巾,—範例性四藝碼器電路包 括四個&quot;選定’’偏壓線與—單—未選定偏壓線。此命名之基 本原理係因為在選擇一給定解碼器頭之輸入(即驅動至一 有效位準)時’該解碼器頭將其輸出耦合至一&quot;選定&quot;偏壓 線。然而’此點絕不暗示著,所示該等頭之全部四個均驅 動其個別輸出至—反映正在選擇該輸出之位準,因為 情況下僅該等選定偏壓線之一係在一適合於一選定輸出之 條件下實際偏壓’而剩餘三個選定偏壓線係在—適合於一 未k定輸出之條件下偏壓。用於一多頭解碼器之該些&quot;選 定•,偏壓線在本文中係說明為一,,源極選擇匯流排”,但類 似操作,另有提醒的除外。某些具體實施例還包括一第二 此類匯流排,其係一&quot;反向源極選擇匯流排&quot;而非一單一未 選定偏壓線。 反之,若用於該多頭解碼器之輸入節點係無效或未選 定,則所有此類頭均驅動其個別輸出至一相關聯,,未選定,1 偏壓線(或一反向源極選擇匯流排之個別匯流排線卜對於 許多有用具體實施例,此類未選定偏壓線可組合成一由該 多頭解碼器之所有頭共用的單一偏壓線。 123178.doc -35- 200826114 類似或相關字線解碼器結構及技術,包括此類解碼之額 外階層、用於該等解碼匯流排(例如XSELN與XSELP)之偏 壓電路阻止及相關支援電路,係進一步說明於R〇y E.The contents of Part (4) are incorporated herein by reference. The dry circuit for such a hierarchical decoder can be used... M Λ1 See US Patent Application Bulletin No. 2006- 0146639 Α1, Luca G Fas〇li牦, • m〇 等h, etc. Apparatus and method for hierarchical decoding of a dense memory array using a multi-level multi-head decoder, the entire disclosure of which is incorporated herein by reference. In the particular material sheet described herein, the exemplary four-yard encoder circuit includes four &quot;select&apos; bias lines and - single-unselected bias lines. The basic principle of this designation is because the decoder head couples its output to a &quot;selected&quot; bias line when selecting the input of a given decoder head (i.e., driving to a valid level). However, 'this does not imply that all four of the heads shown drive their individual outputs to - reflecting the level at which the output is being selected, since only one of the selected bias lines is suitable The actual bias voltage is selected under the condition of a selected output and the remaining three selected bias lines are biased under conditions suitable for a non-k output. The &quot;selected&quot; for a multi-head decoder, the bias line is described herein as a source selection bus, but similar operations, unless otherwise noted. Some embodiments also A second such bus is included, which is a &quot;reverse source selection bus&quot; instead of a single unselected bias line. Conversely, if the input node for the multi-head decoder is invalid or unselected , then all such heads drive their individual outputs to an associated, unselected, 1 bias line (or a separate bus select bus bar individual bus line for many useful embodiments, such The selected bias lines can be combined into a single bias line shared by all of the heads of the multi-head decoder. 123178.doc -35- 200826114 Similar or related word line decoder structures and techniques, including additional levels of such decoding, for The bias circuit blocking and related support circuits of the decoding busses (for example, XSELN and XSELP) are further described in R〇y E.

Scheuerlein及 Matthew P. Crowley之美國專利第 6,856,572 號’標題為”利用具有雙用途驅動器裝置之記憶體陣列線 驅動器之多頭解碼器結構’’,其全部揭示内容以引用形式 併入本文、以及 Roy E · Scheuerlein及 Matthew P. Cro wle 之 美國專利第6,859,410號,標題為”特別適合於介接具有極 小佈局間距之樹狀解碼器結構,,,其全部揭示内容以引用 形式併入本文。 圖12係表示依據本發明之特定具體實施例之一三維記憶 體陣列之一字線層及一位元線層之一俯視圖。其他字線層 及位元線可使用該等所示層來實施並在某些具體實施例中 共用相同的垂直連接。顯示記憶體區塊332、334分別包括 複數個位元線333、335,並具有2:1交錯的字線片段。至 用於一區塊之該等字線片段之一半的垂直連接係在該區塊 左側(例如字線片段337及垂直連接339),而至用於該區塊 之該等字線片段之另一半的垂直連接係在該區塊右側(例 如字線片段336及垂直連接340)。此外,各垂直連接在二 相鄰區塊之各區塊内用於一字線片段。例如,垂直連接 340連接至在陣列區塊332内的字線片段336並連接至陣列 區塊334内的字線片段338。換言之,各垂直連接(例如垂 直連接340)係由在二相鄰區塊之各區塊内的一字線片段來 共用。然而,若所期望的,用於該等第一及最後陣列區塊 123178.doc -36- 200826114 之個別’’外部”垂直連接僅用於該等第一及最後陣列區塊内 的字線片段。例如,若區塊334係形成一記憶體陣列(或一 w己饭體機架)之複數個區塊之最後區塊,則其外部垂直連 接(例如垂直連接344)可僅用於區塊334内的字線片段342, 因而遍及該陣列之其他部分不由二字線片段共用。 藉由父錯所不字線片段,該等垂直連接之間距係個別字 線片#又自身之間距的兩倍。此點特別有利,因為可獲得用 於許多被動元件記憶體單元陣列之字線間距明顯小於可獲 付用於可能用於形成垂直連接之許多通道結構之間距。而 且,此點還可減小字線驅動器電路之複雜性以實施於記憶 體陣列下面的半導體基板内。 現在參考圖13,顯示一示意圖,其表示依據本發明之特 疋具體實施例具有一片段化字線配置之一三維記憶體陣 列。各字線係由在該記憶㈣列之至少一(且較有利的多 個)字線層之一或多個字線片段所形成。例如,一第一字 線係由置放於該記憶體陣列之—字線層上的字線片段36〇 與置放於另一字線層上的字線片段362所形成。該等字線 片段360、362係藉由一垂直連接358來連接以形成第一字 線。垂直連接358還提供一連接路徑至置放於另一層(例如 在該半導體基板内)的驅動器裝置 wz。一來自一列 解碼器(未顯示)之解碼輸出352實質上平行於該等字線片段 360、362而橫過,有時透過裝置172將該等字線片段gw又 ㈣合至一實質上平行於該等字線片段而橫過之又解碼偏 壓線167(例如源極選擇匯流排XSELN),有時透馬妒 123178.doc -37 - 200826114 將4等予線片段360、362耦合至一解碼偏壓線2〇3(例如如 圖9所示之反向源極選擇匯流排XSELp)。 還”、、員示子線片段3 61、3 6 3,其係藉由一垂直連接3 5 9連 接以开成一第二字線並提供一連接路徑至字線驅動器電路 175、176。來自該列解碼器之另一解碼輸出353有時透過 凌置176將该些字線片段36 i、363耦合至解碼源極選擇線 (即偏壓線”)167,有時透過裝置175將該等字線片段361、 363耦合至解碼偏壓線2〇3。儘管此圖示概念性介紹一範例 生陣列組態,但下面說明許多具體實施例,其包括所示組 怨之變更,而且包括可能適用於特定具體實施例,但不一 定適用於全部具體實施例之細節。 在特定較佳具體實施例中,利用一六頭字線驅動器。與 此六頭字線驅動器電路相關聯之六個字線由兩個相鄰記憶 體區塊共用’如在前述美國專利第7,G54,219射所述。換 言之,一給定六頭字線驅動器解碼並驅動二相鄰區塊之各 區塊内的六個字線。如該圖所暗示,該些相鄰區塊可視為 分別在該等相關聯字線驅動器左邊及右邊。然而,在較佳 具體實化W中’ itb類多頭字線驅動器f質上係置放於該等 陣歹】區塊下面,且僅至該等字線之該等垂直連接係製造於 呑亥專區塊之間。 涵蓋具有非鏡射陣列(例如一字線線層僅與一單一位元 線層相關聯)之特定具體實施例,諸如在以⑶α ?“01丨等 人於20G5年3月31申請的美國巾請案第11/G95,9G7號,標題 為&quot;用於在記憶體陣列内併人區塊冗餘之裝置及方法&quot;,現 123178.doc -38 - 200826114 為美國專利第7,142,471號中所述,其全部揭示内容係以引 用形式併入本文。特定言之,圖15顯示4個位元線層、同 時在一陣列區塊之頂部及底部側上的一 16頭行解碼器。此 圖顯示在4位元線層之各層上的4位元線係由一單一 16頭行 解碼器|禺合至頂部資料匯流排(說明4 1/〇層),且同樣地在 相同4位元線層之各層上的4位元線係由一單一 “頭行解碼 器而耦合至該底部資料匯流排(但在該說明中,該等兩個 群組的16選定位元線係位於相同陣列區塊内)。涵蓋其他 半鏡射具體實施例,例如二位元線層共用一字線層之該等 具體實施例,以形成二記憶體陣列。 在接下若干圖中,說明利用重置程式化(即反向偏塵程 式化)之各種具體實施例。因此,一些定義係依次用於本 揭不内容之此部分。術語&quot;設定&quot;應視為正向偏壓一單一(或 群組)記憶體單元,以透過各記憶體單元引起一更低電 阻。術語&quot;抹除&quot;應視為正向偏壓一記憶體單元區塊,以透 過各記憶體單元引起一更低電阻。最後,術語&quot;重置,,應視 為反向偏壓一記憶體單元以透過各此類單元引起一更高電 阻。(關於本文所述之其他具體實施例,此類定義可能不 適用特定。之,術δ吾,,抹除”可還指橫跨一記憶體單元之 反向偏壓條件以增加該單元之電阻。) 外現在參考圖14,&quot;記憶體陣列370包括-第一條371與一 第二條372 °該第—條371係還標註為條G而該第二條372還 心D主為條1。條371包括二記憶體機架ΒΑγ—〇〇與ΒΑγ—〇ι。 各此類記憶體機架包括複數個陣列區塊(例如财類記憶 123178.doc -39- 200826114 體陣列區塊)。儘管顯示此範例性記憶體陣列370包括兩個 記憶體條,各具有兩個記憶體機架,但還涵蓋其他數目的 條及機架。 該第一記憶體機架BAY—00表示其他記憶體機架。總計 表示16個記憶體陣列區塊,其中兩個標註為374及375,各 具有置放於該記憶體陣列下面的一感應放大器(例如在該 等半導體基板層内,但是一或多個記憶體平面可形成於在 該等基板層上形成的一介電層上)。一頂部行解碼器電路 380、一頂部資料匯流排373及一頂部位元線選擇區塊 跨越此機架之丨6個陣列區塊,且與從各陣列區塊之頂部退 出之該等位7G線相關聯。一底部行解碼器電路379、一底 部貧料匯流排378及一底部位元線選擇區塊刊2跨越此機架 之16個陣列區塊,且與從各陣列區塊之底部退出之該等位 元線相關聯。 應明白,頂部行解碼器電路380可說明為在該等陣列區 塊”上面&quot;,而底部行解碼器電路379可說明為在該等陣列 區塊&quot;下面&quot;。此術語視覺上反映示意圖中所示之電路區塊 之方位。此類位置還可描述為在該等陣列區塊&quot;一侧&quot;及 ”相對側”(但此公認地暗示著一水平基板用於其上實施此電 路之積體電路)。此外,方向術語&quot;北&quot;與,,南”係用於說明各 種電路區塊之位置關係的方便術語。 相比之下’在特定具體實施例中,記憶體陣列可形成於 基板”上面而各種電路組塊係說明為在記憶體陣列&quot;下 面&quot;。如本文所適用,在基板或一記憶體陣列區塊(其均係 123178.doc -40- 200826114 一般具有一平面特性之實際實體結構)的,,上面&quot;或&quot;下面&quot;係 相對於-垂直於此類基板或記憶體平面之表面而言。 在圖u中’儘管底部行㈣器可描述為在該等陣列區塊 &quot;下面&quot;,但此類行解石馬器不—定在該記憶體陣列下面(即更 罪近基板)。減之下’可假定描料在該陣列區塊邊界 内並描述為在該陣列區塊&quot;下面”或&quot;之下”的該等感應放大 盗區塊(標註為SA),以傳遞此類實體位置及結構關係。在U.S. Patent No. 6,856,572 to Scheuerlein and Matthew P. Crowley, entitled "Multi-head Decoder Structure Using a Memory Array Line Driver with Dual-Purpose Driver Devices", the entire disclosure of which is incorporated herein by reference, U.S. Patent No. 6,859,410 to Scheuerlein and Matthew P. Crow, entitled "Specially Suitable for Interconnecting Tree-Decoder Structures with Very Small Layout Spacing," the entire disclosure of which is incorporated herein by reference. Figure 12 is a top plan view showing one of a word line layer and a bit line layer of a three-dimensional memory array in accordance with a particular embodiment of the present invention. Other word line layers and bit lines can be implemented using the layers shown and share the same vertical connections in some embodiments. Display memory blocks 332, 334 include a plurality of bit lines 333, 335, respectively, and have 2: 1 interleaved word line segments. The vertical connection to one half of the word line segments for a block is to the left of the block (e.g., word line segment 337 and vertical connection 339), and to the other word line segments for the block Half of the vertical connections are to the right of the block (eg, word line segment 336 and vertical connection 340). In addition, each vertical connection is used in a block of two adjacent blocks for a word line segment. For example, vertical connection 340 is coupled to word line segment 336 within array block 332 and to word line segment 338 within array block 334. In other words, each vertical connection (e.g., vertical connection 340) is shared by a word line segment within each of the two adjacent blocks. However, if desired, the individual ''outside' vertical connections for the first and last array blocks 123178.doc-36-200826114 are only used for word line segments within the first and last array blocks. For example, if block 334 forms the last block of a plurality of blocks of a memory array (or a self-contained rack), then its external vertical connection (eg, vertical connection 344) may be used only for the block. The word line segment 342 within 334 is thus not shared by the second word line segments throughout the other portion of the array. By the parent fault, the vertical line is separated by two lines between the individual word lines and the self. This is particularly advantageous because the word line spacing available for many passive element memory cell arrays is significantly less than the distance between many channel structures that can be used for forming vertical connections. Moreover, this can also be reduced. The complexity of the small word line driver circuit is implemented in a semiconductor substrate underneath the memory array. Referring now to Figure 13, a schematic diagram is shown showing a fragmented word in accordance with a particular embodiment of the present invention. Configuring one of the three-dimensional memory arrays. Each word line is formed by one or more word line segments of at least one (and advantageously a plurality of) word line layers in the memory (four) column. For example, a first word line The word line segment 36〇 placed on the word line layer of the memory array and the word line segment 362 placed on the other word line layer are formed by the word line segments 360 and 362. A vertical connection 358 is connected to form a first word line. The vertical connection 358 also provides a connection path to a driver device wz placed in another layer (e.g., within the semiconductor substrate). A column of decoders (not shown) The decoded output 352 traverses substantially parallel to the word line segments 360, 362, and sometimes the device 172 combines the word line segments gw (4) to a substantially parallel slant of the word line segments. Decoding bias line 167 (e.g., source select bus XSELN), sometimes translating 4, etc. pre-wire segments 360, 362 to a decode bias line 2〇3 (eg, as shown The reverse source selection bus (XSELp) shown in Figure 9. Also, the member shows the sub-line 3 61,3 63, which line is connected by a vertical connection 359 to open a second word line and to provide a connection path to the word line driver circuits 175 and 176. Another decoded output 353 from the column decoder is sometimes coupled to the decoded source select line (i.e., bias line) 167 via the ping 176, which is sometimes transmitted by the device 175. The equal word line segments 361, 363 are coupled to the decode bias line 2 〇 3. Although this illustration conceptually describes a sample live array configuration, many specific embodiments are described below, including variations of the illustrated grievances, and include It may be suitable for a particular embodiment, but not necessarily for the details of all embodiments. In a particularly preferred embodiment, a six-word line driver is utilized. Six word lines associated with the six-word line driver circuit Shared by two adjacent memory blocks as described in the aforementioned U.S. Patent No. 7, G54, 219. In other words, a given six-word line driver decodes and drives six of the blocks in two adjacent blocks. Word lines. As the figure implies, the adjacent blocks can be considered to be on the left and right sides of the associated word line drivers respectively. However, in a better specific implementation, the 'itb class multi-word line driver f quality The upper system is placed in these垂直] Below the block, and only the vertical connections of the word lines are fabricated between the blocks of the Haihai area. Covering with non-mirror arrays (eg, a word line layer is only associated with a single bit line layer) Specific embodiments of the invention, such as the US towel application No. 11/G95, 9G7 filed on March 31, 20G, by (3) α? "01丨 et al., entitled "&quot; for use in a memory array and The apparatus and method for the redundancy of a human block is described in U.S. Patent No. 7,142,471, the entire disclosure of which is incorporated herein by reference. In particular, Figure 15 shows a 4-bit line layer with a 16-bit row decoder on top and bottom sides of an array block. This figure shows that the 4-bit line on each layer of the 4-bit line layer is combined by a single 16-head row decoder to the top data bus (description 4 1/〇 layer), and is equally in the same 4 bits. The 4-bit line on each layer of the meta-layer is coupled to the bottom data bus by a single "head-to-line decoder" (but in this description, the 16 selected positioning elements of the two groups are located at the same Within the array block, other specific embodiments are contemplated, such as a two-bit line layer sharing a word line layer, to form a two-memory array. In the following figures, the use of weight is illustrated. Various specific embodiments of stylized (ie, reverse dusting stylization). Therefore, some definitions are used in turn for this part of the content. The term &quot;setting&quot; should be considered as a single forward bias ( Or group of memory cells to cause a lower resistance through each memory cell. The term &quot;erase&quot; should be considered as a forward biasing of a memory cell block to cause a change through each memory cell. Low resistance. Finally, the term &quot;reset, should be considered as a Biasing a memory cell to cause a higher resistance through each such cell. (For other embodiments described herein, such definitions may not apply to a particular one. The reverse bias condition across a memory cell increases the resistance of the cell.) Referring now to Figure 14, the &quot;memory array 370 includes a first strip 371 and a second strip 372 ° the strip 371 The system is also labeled as the strip G and the second strip 372 is also the center of the main D. Strip 371 includes two memory racks ΒΑγ-〇〇 and ΒΑγ-〇ι. Each such memory rack includes a plurality of array blocks (e.g., financial memory blocks 123178.doc - 39 - 200826114 body array blocks). Although this exemplary memory array 370 is shown to include two memory strips, each having two memory racks, other numbers of strips and racks are also contemplated. The first memory rack BAY_00 represents another memory rack. A total of 16 memory array blocks, two of which are labeled 374 and 375, each having a sense amplifier placed under the memory array (eg, within the layers of the semiconductor substrate, but one or more memories) A plane can be formed on a dielectric layer formed on the substrate layers). A top row decoder circuit 380, a top data bus 373 and a top bit line selection block span the six array blocks of the rack and exit the same 7G from the top of each array block Line associated. A bottom row decoder circuit 379, a bottom lean bus 378 and a bottom bit line selection block 2 spanning 16 array blocks of the rack and exiting from the bottom of each array block The bit line is associated. It will be appreciated that the top row decoder circuit 380 can be illustrated as being "above" the array blocks, and the bottom row decoder circuit 379 can be illustrated as being "in the &quot;below&quot;. The orientation of the circuit blocks shown in the schematic diagram. Such locations may also be described as &quot;one side&quot; and &quot;opposite side&quot; in the array blocks (but this is generally acknowledged to be a horizontal substrate for implementation thereon) In addition, the directional term &quot;North&quot; and, "South" is a convenient term used to describe the positional relationship of various circuit blocks. In contrast, 'in a particular embodiment, a memory array can be formed on a substrate" and various circuit blocks are illustrated as being in a memory array &quot;below&quot; as applied herein, on a substrate or a memory Array blocks (both of which are 123178.doc -40-200826114 generally having a planar physical property), above &quot;or &quot;lower&quot; relative to - perpendicular to such substrate or memory plane In terms of the surface, in Figure u, 'Although the bottom row (four) can be described as "below" in the array block, such a stone is not fixed under the memory array (ie, more sin Near-substrate). Subtracting 'can be assumed to be within the boundaries of the array block and described as under the array block "below" or "below" the inductively amplified block (labeled SA) To convey the location and structural relationship of such entities.

/ 本說明書及各種圖之背景下,應清楚&quot;上面&quot;及&quot;下面&quot;之 法。 在特定範例性具體實施财,該等位元線解碼器係⑹貝 解碼器,ϋ同時選擇在-敎記憶體陣列區塊之頂部側上 的16位元線。此&quot;選擇&quot;涉及行解碼,不一定暗示著所有16 位元線實際上同時程式化。該等十六個選定位元線較佳的 係在四個位元線層之各層上配置成在頂部(或用於其他解 碼器之底部)退出該陣列的四個相鄰位元線。 頂部資料匯流排373之該等十六個I/O線水平橫越所有十 六個區塊。此類匯流排對應於上述SEL]B匯流排。此資料 匯流排373之該等個別匯流排線之各匯流排線係耦合至分 佈於該等所示十六個區塊中之十六個感應放大器電路之_ 個別者。该等十六個資料匯流排線之各資料匯流排線還可 輕合至一相關聯偏壓電路(即一重置電路),其可在一特定 操作模式期間用於適當偏壓該等”選定”16個位元線内的該 等個別位元線。例如,對於一重置操作模式,此重置電路 依據用於該等16位元線之各位元線之資料位元,並還依據 123178.doc -41 - 200826114 允許同時程式化之纟元線數目(當然意味著耦纟至特定位 元線之欲程式化單元)來適當偏壓該等&quot;選定&quot;16位元線内的 該等欲程式化位元線與該等不欲程式化位元線。可停用該 些偏壓電路並引起其在該等選定位元線藉助資料匯流排 3 73(即上述SELB匯流排)而耦合至個別感應放大器時在一 讀取操作模式期間展現一高阻抗。 底部資料匯流排378之該等十六個1/〇線水平橫越所有十 六個區塊。此類匯流排對應於上述另_SELB匯流排,此 時係用於在陣列底部退出之該等位元線(應記住該等位元 線係2:1交錯)。如前述,&amp;資料匯流排爪之該等個別匯流 排線之各匯流排線係耦合至分佈於該等所示十六個區塊中 的十六個感應放大器電路之一個別者。在各群組的16區塊 =一機架)中,存在連接至32個選定位元的32個感應放大 器。在讀取模&lt;了,所有該等選定位元線可配置成用以落 入該等十六個區塊之-内,或可另外配置,如此處將要說 明。该等感應放大器可方便地實施於記憶體陣列區塊之 下,但該等資料匯流排線373、378、該等十六頭行選擇解 碼器(即該等位元線選擇區塊381、382)及該等行解碼器 379、380之一小部分較佳的係實施於該陣列區塊外部。有 用行解碼器配置之額外細節可見諸於美國申請案第 1 1/095,907號(美國專利第7,142,471號)以及前述美國專利 申請公告案第2006-0146639 A1號中。 在一程式化模式下,總程式化電路之數量可限制同時程 式化記憶體單元之數目。此外,沿一單一選定位元線或字 123178.doc -42- 200826114 =之程,流之數量還可限制記憶體單元之數目, ,、可同時可靠地加以程式化。々 各行鰓&amp; _ 不靶例性采構中,若兩 丁解馬嗔相同陣綱内的位元線, 總計選定32個位元線。蚊各 ^攸疋谷解碼器從四個位元線層之各 層中、擇四個位元線(即來自各個別記憶體平面之四個位 儿線)’則在各記憶體平面上的選定字線片段須支援用於 總计八個選定記憶體單元之mm見圖13以顯 :每層的個別字線片段。)該些選定記憶體單元之四個記 :體:元與向北退出之位元線相關聯,而其他四個選定記 fe體單7L係與向南退出位元線相關聯。該等選定記憶體單 元之全部32個單元將由相同字線驅動器電路來驅動,但是 該等選定記憶體單元之各記憶體單元係由其自身位元線驅 動器電路來驅動。/ In the context of this manual and the various diagrams, it should be clear that the above &quot;and&quot; In a particular exemplary implementation, the bit line decoder is a (6) decoder that simultaneously selects a 16-bit line on the top side of the --memory array block. This &quot;select&quot; relates to row decoding, not necessarily implying that all 16-bit lines are actually stylized at the same time. Preferably, the sixteen selected bit lines are arranged on each of the four bit line layers to exit the four adjacent bit lines of the array at the top (or at the bottom of other decoders). The sixteen I/O lines of the top data bus 373 traverse all of the sixteen blocks horizontally. Such a bus bar corresponds to the above-mentioned SEL]B bus bar. The bus lines of the individual bus bars of the data bus 373 are coupled to the individual of the sixteen sense amplifier circuits distributed among the sixteen blocks shown. The data bus bars of the sixteen data bus bars can also be lightly coupled to an associated bias circuit (ie, a reset circuit) that can be used to properly bias such a mode of operation during a particular mode of operation. "Select" the individual bit lines within the 16 bit lines. For example, for a reset mode of operation, the reset circuit is based on the data bits for the bit lines of the 16-bit lines, and also allows the number of simultaneous programming lines according to 123178.doc -41 - 200826114 (Of course means a stylized unit coupled to a particular bit line) to properly bias the &apos;select&quot;16 bit lines of the desired stylized bit line and the undesired stylized bits Yuan line. The bias circuits can be deactivated and cause them to exhibit a high impedance during a read mode of operation when the selected bit lines are coupled to the individual sense amplifiers via data bus 3 73 (ie, the SELB bus) . The sixteen 1/〇 lines of the bottom data bus 378 traverse all of the sixteen blocks horizontally. Such busbars correspond to the other SELB busbars described above, which are used to exit the bitlines at the bottom of the array (it should be remembered that the bitlines are 2:1 interleaved). As previously described, each of the individual bus bars of the &amp; data sinking pawl is coupled to one of the sixteen sense amplifier circuits distributed among the sixteen blocks shown. In each group of 16 blocks = one rack, there are 32 inductive amplifiers connected to 32 selected positioning elements. In reading the mode &lt;, all of the selected positioning elements can be configured to fall within the sixteen blocks, or can be otherwise configured, as will be explained herein. The sense amplifiers can be conveniently implemented under the memory array block, but the data bus lines 373, 378, the sixteen header select decoders (ie, the bit line select blocks 381, 382) And a small portion of the row decoders 379, 380 are preferably implemented external to the array block. Additional details of the use of a row decoder configuration can be found in U.S. Patent Application Serial No. 1 1/095,907 (U.S. Patent No. 7,142,471), and the aforementioned U.S. Patent Application Publication No. 2006-0146639 A1. In a stylized mode, the total number of stylized circuits can limit the number of simultaneous memory cells. In addition, along a single selected location line or word 123178.doc -42- 200826114 =, the number of streams can also limit the number of memory cells, and can be reliably stylized at the same time. 々 In each row &amp; _ non-targeted structure, if two bits of the same line in the same array are solved, a total of 32 bit lines are selected. The mosquitoes each selects four bit lines (ie, four bit lines from the respective memory planes) in each of the four bit line layers to select on each memory plane. The word line segment shall support mm for a total of eight selected memory cells as shown in Figure 13 to show: individual word line segments for each layer. The four memory cells of the selected memory cells are associated with a bit line that exits north, and the other four selected cells are associated with a southbound exit bit line. All 32 cells of the selected memory cells will be driven by the same word line driver circuit, but each memory cell of the selected memory cells is driven by its own bit line driver circuit.

如上所暗示,即便用於32個單元之總程式化電流可由該 積體電路來供應,用於8個選定記憶體單元之程式化電流 可沿各層上的該等選定字線片段引起一不可接收的電壓 降。此外’該選定字線驅動器電路可能無法使用可接收電 壓降來驅動此類電流。 在重置程式化模式下,將一反向偏壓施加至各選定被 動記憶體單元,藉此將可修改的電阻材料重置至一高電阻 狀態以程式化使用者資料。在一區塊内的一或多個位元線 可選定用於同時程式化,且隨著該等位元之某些位元重置 至一更高電阻狀態,從該選定位元線流向該選定字線之電 流明顯減小,且該等剩餘位元由於減小的字線汉降而看見 123178.doc -43- 200826114 ’更容易程式化之該等位元先改 的位元看見一略微更高的電壓以 一明顯更高的電壓。由此, 變狀態,從而使更’’頑固’,的 幫助程式化此類位元。 雖然如此, 但使32個選定記憶體單元駐留於相As suggested above, even though the total programmed current for 32 cells can be supplied by the integrated circuit, the programmed current for the eight selected memory cells can cause an unacceptable along the selected word line segments on each layer. The voltage drop. In addition, the selected word line driver circuit may not be able to use a receivable voltage drop to drive such current. In the reset stylized mode, a reverse bias is applied to each of the selected passive memory cells, thereby resetting the modifiable resistive material to a high resistance state to program the user data. One or more bit lines within a block may be selected for simultaneous programming, and flow to the selected bit line as the bits of the bit are reset to a higher resistance state The current of the selected word line is significantly reduced, and the remaining bits are seen due to the reduced word line drop. 123178.doc -43- 200826114 'It is easier to stylize the bit changed first. Higher voltages are at a significantly higher voltage. Thus, the state is changed, so that the help of the more ''stubborn'' is stylized. Nonetheless, but 32 selected memory cells reside in the phase

之一個別者。在圖中 同陣列區 原因而難以接受。因此,兩個不同 式化’各使用該等兩個資料匯流排 陣列區塊374係交叉陰影線繪製以 表示其用於重置程式化之選擇。用於區塊374之該等頂部 行解碼器380輸出之一係有效,從而將16個選定位元線耦 合至頂部資料匯流排373(由從陣列區塊374至資料匯流排 373之箭頭表示)。此外,陣列區塊375係交叉陰影線繪製 以表不其用於重置程式化之選擇。用於區塊375之該等底 部行解碼器379輸出之一係也有效,從而將16個選定位元 線柄合至底部資料匯流排378(由從陣列區塊375至資料匯 流排3 7 8之前頭表示)。 一單一列377係由在該記憶體陣列之任一侧上的該等全 域列解碼器(未顯示)來選定,其橫跨整個條371來驅動一全 域列選擇線。此類全域列選擇線對應於圖9所示之列解碼 器電路之解碼輸出158。一多頭字線驅動器電路係致能(藉 由在其源極選擇匯流排及反向源極選擇匯流排上的適當偏 壓條件)以驅動區塊374内的一選定字線376與在區塊375内 的一選定字線。由於共用在此範例性具體實施例内的該等 字線’故一此類選定字線驅動器電路驅動在二區塊374、 3 7 5内的字線。整個程式化電流仍透過此一選定字線驅動 123178.doc -44- 200826114 器電路而發起,但現在沿各選定字線片段之電流減半,由 於各字線片段現在僅支持4個選定記憶體單元。應注意, 在區塊374及375内的下—更高或更低字線係藉由二分離字 線驅動器裝置來驅動且在該等字線驅動器裝置之任一者内 的峰值電流將大約為一半。藉由選擇以將資料頁配置於一 對應於奇或偶字線之更複雜區塊配置内,可完全避免共用 字線驅動器。例如,假定從—給料列區塊左側驅動偶字 線,且從一給定陣列區塊右側驅動奇字線。當在給定陣列 區塊内選定一偶字線時,可同時選定其左邊的區塊,且當 在、々疋陣列區塊内選擇一奇字線時,可同時選定其右邊 的區塊。在此情況下,沒有任何選定字線出現在一未選定 陣列區塊内。在一替代性具體實施例中,欲寫入資料頁可 配置成用以避免共用字線驅動器。 在上述雙資料匯流排範例中,各記憶體區塊係與兩個資 料匯流排373、378相關聯。在一不同記憶體循環中,與陣 列區塊374相關聯之其他位元線將被耦合至底部資料匯流 排378,而與陣列區塊375相關聯之其他位元線將被耦合至 頂部資料匯流排373。在此及其他具體實施例中,為了最 佳化效能,在一給定機架内選定用於讀取之該等區塊不同 於選定用於重置之該等區塊。一次選定一單一區塊用於讀 取,但選定兩個區塊用於重置。該等兩個資料匯流排二者 均讀取有效,但存取一單一區塊,不同於上述重置存取。 存在提供類似好處之各種其他雙資料匯流排配置。圖i 5 顯示一 §己憶體機架400,其中該等奇數記憶體區塊係僅與 123178.doc -45- 200826114 一第一資料匯流排相關聯,而該等偶數記憶體區塊係僅與 一第二資料匯流排相關聯。奇數陣列區塊4〇6係與該第一 資料匯流排402相關聯,其係表示為位元線選擇區塊4〇8, 而偶數陣列區塊407係與第二資料匯流排4〇4相關聯。二記 憶體陣列區塊(例如陣列區塊406、4〇7)係同時選定,各將 其選定位元線耦合至該等資料匯流排(表示為個別粗體箭 頭 410、412)之一。 圖16顯示一記憶體機架420,其中各記憶體區塊係與一 第一資料匯流排422與一第二資料匯流排424二者相關聯。 在一所不記憶體循環中,第一陣列區塊426係選定並將其 選定位兀線耦合(粗體箭頭430)至第一資料匯流排422,而 第二陣列區塊427係同時選定並將其選定位元線耦合(粗體 箭頭432)至第二資料匯流排424。在另一記憶體循環中, 第一陣列區塊426可被選定並將其選定位元線耦合至第二 貧料匯流排424,而第二陣列區塊427係同時選定並將其選 定位元線搞合至第一資料匯流排422。 圖17顯示一記憶體機架44〇,其中各記憶體區塊係與一 第一資料匯流排442與一第二資料匯流排4料二者相關聯, 其均位於该等陣列區塊之相同側。第一陣列區塊446係憑 藉一第一位元線選擇區塊449而與該第一資料匯流排442相 關聯,並還憑藉一第二位元線選擇區塊448而與第二資料 匯流排444相關聯。在所示範例性記憶體循環中,二記憶 體陣列區塊(例如陣列區塊447、446)係同時選定,各將其 選定位元線耦合至該等第一及第二資料匯流排料2、 123178.doc -46- 200826114 444(表示為個別粗體箭頭450、454)。 現在參考圖18,描述一記憶體機架460,其類似於上述 記憶體機架BAY一00,除了在此範例性具體實施例中,兩 個同時選定陣列區塊462、464係不相鄰。在一所示記憶體 循環中,陣列區塊462係選定並將其選定位元線耦合(即粗 體箭頭)至上部資料匯流排466,而陣列區塊464係同時選 定並將其選定位元線耦合至一下部資料匯流排468。此組 織在相鄰記憶體陣列區塊之間不共用該等字線時特別有 用,但即便共用此類字線之情況下仍能使用。在此情況 下,在一選疋區塊内的一選定字線還將伸入相鄰記憶體區 塊内。 在該些所示具體實施例之各具體實施例中,一個以上區 塊係選定用於重置程式化。反向偏壓係施加至該等選定陣 列區塊(即選定&quot;子陣列”)内的該等被動元件單元,藉此將 可修改電阻材料重置至一高電阻狀態以將使用者資料程式 化在該陣列内。此可能由於若干原因在高帶寬下完成。首 先,藉由選擇一個以上區塊用於程式化,可增加同時程式 化記憶體單元之數目超出一給定字線片段所強加或甚至一 給疋子線驅動器電路所強加之該等限制。可選定兩個以上 選定陣列區塊,只要該等資料匯流排到達各此類區塊。此 外,該程式化方向有助於允許程式化更大數目的單元。換 言之,由於該等程式化位元之某些位元重置至一更高電阻 狀悲,從位兀線流向字線之電流數量明顯下降,故勝餘位 元由於不斷減小的字線電壓降而看見略微更高的電壓。對 123178.doc -47- 200826114 於一給定最大程式化電流’可能從低至高電阻比從高至低 電阻可靠地程式化更多位元。還貢獻於一高帶寬程式化的 係在全部大量未選定字線及位元線上的該等偏壓條件。由 於S亥些線全部保持接地,故不存在選定及取消選定陣列區 塊時升壓偏壓該等未選定陣列陣列相關聯之較大延遲,也 不存在必須谷納以上升及下降偏麼此類陣列區塊之較大電 流暫態電流。應注意,在此重置程式化配置中,甚至在選 定記憶體區塊内的該等未選定字線與位元線偏壓在接地下 (即使用特定範例性解碼器結構時向左浮動)。 在範例性具體實施例中,K織一記憶體晶Μ,使得各 機架具有其自己的讀取寫入電路組與將該等讀取/讀寫電 路連接至位元線選擇電路之至少—資料匯流排。此匯流排 橫跨該機架之寬度而延伸,或換言之&quot;跨越,,區塊群組。可 能存,在該等區塊頂側的—行解碼器以及在該等區塊底側 的-第二行解碼器,故存在兩個資料匯流排。在特定具體 實施例中,可能存在與各個別f料匯流排相關聯的兩組讀 :寫入電路。較佳的係一特定資料頁係分散至所有機架以 獲得最高帶寬。此點係藉由在各記憶體機架内的__對選定 陣列區塊而描述於圖14所示之範例性具體實施例中。 較佳的係該等選定位亓孫八士 、疋位兀係分佈於一機架内的兩個區塊 ’一區塊具有由該等行解碼器之—敎並於該等資料匯 &gt;4之-相關聯之位元線’而第二區塊係由另一行解以 及貪《流排來選定,使得每機架加倍帶寬,但在任一字 、本片&amp;内·動之電流不變。此外,在一選定行位置之該等 123178.doc -48- 200826114 位元線之-或許多者係同時選定用於重置程式化。同時程 j化的數目可能受從一區塊内的該等選定位元線流向共用 字線之電流的限制。但此限制在一方法中得到減輕,其中 由於該等位元之某些位元重置至一更高電阻狀態,透過 6重置&quot;單70之電流減小,沿共用字線片段之IR降減小, 且剩餘位元獲得更多電壓以促進其重置。 在各選定區塊㈣料敎字線較佳的係全部在相同列 ^ 上,從而消除解碼蘊涵,因為該全域列解碼器電流不需要 k化以支援此點。較佳的係該等同時選定區塊係相鄰,特 : 別在相鄰區塊之間共用字線之情況下。該解碼可配置,使 • 得對於在兩個相鄰區塊之間共用的任一選定字線,該些兩 個相鄰陣列區塊可經組態成同時選定陣列區塊。例如,置 放於該等及第二區塊之間的一給定字線驅動器驅動在該第 一及第二區塊(二者同時選定)内的一共用字線。下一字線 (假定其係在該等陣列區塊之左右側採用2:丨交錯形式)將由 、該等第二及第三陣列區塊之間的一陣列線驅動器來驅動, 该等陣列區塊還可以係選定陣列區塊。此點避免處理選定 字線伸入相鄰非選擇陣列區塊内。 當使用重置程式化時,各記憶體單元係藉由該”設定,,操 . 作模式而設定回至一低電阻狀態,該設定操作模式可用於 再寫新資料,或藉由一次施加正向偏壓至一位元,咬在一 資料頁或一抹除區塊内的許多位元來抹除一群組位元。高 效成抹除可藉由在一區塊内選擇多個位元線或多個字線, 並將該專早元设疋至低電阻來獲得。在位元驅動器路押内 123178.doc -49- 200826114 的限流電路限制流動至共用字線之總電流。取決於所選記 憶體單元技術、及設定電流及重置電流之相對數量、及u 單元洩漏電流之數量,可比用於重置(即程式化)選定更少 區塊用於設定或抹除操作。 一電阻材料之選擇係形成二極體之多晶矽材料。一反熔 絲(’AF’’)可與多晶矽二極體串列,且該反熔絲係在製造時 一格式化步驟中在程式化事件之前跳變。該反熔絲用於限 制設定時單元傳導之最大電流。 如上述,較佳的係該記憶體陣列包括一片段化字線架構 (如圖12及13所示),且較佳的係一三維陣列。在特定具體 實施例中,在一給定字線層上的該等字線係與在一單一位 元線層上的位元線相關聯,而在特定具體實施例中,在一 所謂”半鏡射”配置中’在-給定字線層上的該等字線在二 位元線層(即定義二記憶體平面的一單一字線層與二位元 線層)之間共用。此類記憶體陣列結構進一步說明於前述 美國專利第6,879,505號中。 至此該等各種解碼器電路之說明主要集中於說明一單一 陣列區塊。應記得,各解碼器曾在_源極選擇匯流排以及 對於某些具體實施例-反向源極選擇匯流排之背景下加以 說明。該字線解碼階層可視為相對直接。源匯流排 及未敎㈣線或者反向雜選擇匯流㈣基於位址資訊 來解碼,並依據哪個陣列區塊係有效來_。本文中他處 已參考類似列解碼電路。可使用於與未衫陣列區塊相關 聯之字線的該(等)個別源極選擇匯流排及/或未選定偏麼線 123178.doc •50- 200826114 向左浮動。 關於該等行解碼器配置,一階層式匯流排配置可用於提 供讀取/寫入資料之有效路由以及在選定及未選定陣列區 塊内之位元線之有效偏壓。將在圖9及10所示之雙源極選 擇匯流排解碼器之背景下說明有用階層式匯流排配置,但 該些配置可調適用於其他解碼器具體實施例。 在u亥#正向操作(讀取及設定)中,一範例性階層式匯流 排配置在該SELN匯流排上提供一適當偏壓用於一選定陣 列區塊,並使未選定陣列區塊的SELN匯流排浮動。此點 有助於減小相鄰一選定陣列區塊内之陣列區塊内的不合需 要功率消耗。在一選定陣列區塊内的該等未選定字線係偏 壓在一相當高電壓VUX(例如VPP_VT)下,且在一共用字線 架構下,該些未選定字線還延伸至相鄰未選定陣列區塊 (即在未選定陣列區塊内的該等字線的一半與該選定陣列 區塊一起共用)。在相鄰陣列區塊内的該等未選定位元線 較佳的係在該未選定位元線電壓VUB(例如ντ)下偏壓。此 點由於透過未選定記憶體單元之洩漏電流而浪費功率。在 相鄰未選定陣列區塊内的該等字線之另—半係浮動,使得 其洩漏直至VUB電壓,故洩漏功率係針對該等未選定單元 之一半而最小化。 該範例性階層式匯流排配置還在—重置操作模式下提供 車又長SELN.路’其跨越許多區塊以到達在該等陣列區 塊下面分佈的該等重置資料驅動器。 四個範例性階層式匯流排配置係描述於接下四圖中。現 123178.doc 200826114 在參考圖19,描述一匯流排配置500,其包括三個記憶體 陣列區塊502、504、506,其表示在一機架内的所有陣列 區塊。儘管僅顯示三個陣列區塊,但應清楚該配置之遞增 性質以及其至任一陣列區塊數目的延伸性。顯示用於各個 別陣列區塊的-個別SELN匯流排片段。如本文所使用, 一匯流排片段僅係比其他此類匯流排更小的一匯流排,而 在其他具體實施例(下述)中,多個匯流排片段可一起耦合One of the individual. It is unacceptable in the same array area in the figure. Thus, the two differentizations each use the two data bus array blocks 374 to be cross-hatched to indicate their choice for resetting the stylization. One of the outputs of the top row decoder 380 for block 374 is active, thereby coupling the 16 selected positioning element lines to the top data bus 373 (represented by the arrows from array block 374 to data bus 373) . In addition, array block 375 is cross-hatched to indicate that it is used to reset the stylized selection. One of the output of the bottom row decoder 379 for block 375 is also valid, thereby consolidating the 16 selected positioning line handles to the bottom data bus 378 (from the array block 375 to the data bus 3 7 8 Before the head indicates). A single column 377 is selected by the global column decoders (not shown) on either side of the memory array, which drives a global column select line across the entire strip 371. Such a global column select line corresponds to the decode output 158 of the column decoder circuit shown in FIG. A multi-word line driver circuit is enabled (by appropriate bias conditions on its source select bus and reverse source select bus) to drive a selected word line 376 and region within block 374 A selected word line within block 375. Since such word lines are shared within this exemplary embodiment, such a selected word line driver circuit drives word lines within two blocks 374, 375. The entire programmed current is still initiated by this selected word line driver 123178.doc -44- 200826114 circuit, but now the current along each selected word line segment is halved, since each word line segment now only supports 4 selected memories unit. It should be noted that the lower-higher or lower word lines within blocks 374 and 375 are driven by the two separate word line driver devices and the peak current in either of the word line driver devices will be approximately half. The shared word line driver can be completely avoided by selecting to configure the data page in a more complex block configuration corresponding to odd or even word lines. For example, assume that the even word line is driven from the left side of the -feed column block and the odd word lines are driven from the right side of a given array block. When an even word line is selected within a given array block, the block to the left can be selected at the same time, and when an odd word line is selected in the array block, the block to the right can be selected at the same time. In this case, no selected word lines appear in an unselected array block. In an alternate embodiment, the data page to be written can be configured to avoid sharing the word line driver. In the dual data bus example described above, each memory block is associated with two data bus bars 373, 378. In a different memory cycle, other bit lines associated with array block 374 will be coupled to bottom data bus 378, while other bit lines associated with array block 375 will be coupled to the top data bus 373. In this and other embodiments, for optimal performance, the blocks selected for reading in a given rack are different than the blocks selected for reset. A single block is selected for reading at a time, but two blocks are selected for reset. Both data busses are valid for reading, but accessing a single block is different from resetting the above. There are various other dual data bus configurations that offer similar benefits. Figure i5 shows a § reciprocal rack 400, wherein the odd-numbered memory blocks are only associated with a first data bus of 123178.doc -45-200826114, and the even-numbered memory blocks are only Associated with a second data bus. The odd array block 4〇6 is associated with the first data bus bar 402, which is represented as a bit line selection block 4〇8, and the even array block block 407 is associated with the second data bus bar 4〇4. Union. The two memory array blocks (e.g., array blocks 406, 4〇7) are simultaneously selected, each coupling its selected location line to one of the data bus bars (denoted as individual bold arrows 410, 412). Figure 16 shows a memory rack 420 in which each memory block is associated with a first data bus 422 and a second data bus 424. In a non-memory cycle, the first array block 426 is selected and positioned to coordinate the line coupling (bold arrow 430) to the first data bus 422, while the second array block 427 is simultaneously selected and The selected location line is coupled (bold arrow 432) to the second data bus 424. In another memory cycle, the first array block 426 can be selected and its selected location line coupled to the second lean bus 424, while the second array block 427 is simultaneously selected and positioned to locate the line. Engaged in the first data bus 422. Figure 17 shows a memory frame 44A, wherein each memory block is associated with a first data bus 442 and a second data bus 4, all of which are located in the same array block. side. The first array block 446 is associated with the first data bus 442 by a first bit line selection block 449, and is also associated with the second data bus by a second bit line selection block 448. 444 is associated. In the exemplary memory cycle illustrated, two memory array blocks (e.g., array blocks 447, 446) are simultaneously selected, each coupling its selected location line to the first and second data sinks. 123178.doc -46- 200826114 444 (expressed as individual bold arrows 450, 454). Referring now to Figure 18, a memory bay 460 is depicted which is similar to the memory bay BAY-00 described above except that in this exemplary embodiment, two simultaneously selected array blocks 462, 464 are not adjacent. In a memory cycle, array block 462 is selected and selected for location line coupling (ie, bold arrows) to upper data bus 466, while array block 464 is simultaneously selected and positioned to locate the line. Coupling to the lower data bus 468. This organization is particularly useful when the word lines are not shared between adjacent memory array blocks, but can be used even if such word lines are shared. In this case, a selected word line within a selected block will also extend into the adjacent memory block. In the specific embodiments of the illustrated embodiments, more than one block is selected for resetting the stylization. A reverse bias is applied to the passive component cells within the selected array block (ie, selected &quot;sub-array)), thereby resetting the modifiable resistive material to a high resistance state to program the user data Within the array. This may be done at high bandwidth for several reasons. First, by selecting more than one block for stylization, the number of simultaneous stylized memory cells can be increased beyond that imposed by a given wordline segment. Or even impose such restrictions on the scorpion driver circuit. More than two selected array blocks can be selected as long as the data bus reaches each such block. In addition, the stylized direction helps to allow the program. A larger number of cells. In other words, since some of the bits of the stylized bits are reset to a higher resistance, the amount of current flowing from the bit line to the word line is significantly reduced, so the winning bit is due to Decreasing word line voltage drop and seeing a slightly higher voltage. For 123178.doc -47- 200826114 at a given maximum stylized current 'may be from low to high resistance ratio from high to low resistance reliable Stylizes more bits. It also contributes to a high-bandwidth stylized set of bias conditions on all of the large number of unselected word lines and bit lines. Since all of the lines are kept grounded, there is no selection and cancellation. When the array block is selected, the boost bias voltage has a large delay associated with the unselected array arrays, and there is no large current transient current that must be used to rise and fall in such array blocks. In this reset stylized configuration, even the unselected word lines and bit line biases in the selected memory block are grounded (ie, float to the left when using a particular exemplary decoder structure). In a specific embodiment, K is a memory wafer such that each rack has its own read write circuit group and at least the data read/write circuit is connected to the bit line selection circuit. Row. This busbar extends across the width of the rack, or in other words, "crossing," a group of blocks. It may be stored on the top side of the block - the row decoder and the bottom side of the block - the second line of decoders, so there are two Data bus. In a particular embodiment, there may be two sets of read: write circuits associated with respective bus bucks. Preferably, a particular data page is spread across all racks to obtain the highest bandwidth. This point is described in the exemplary embodiment shown in Figure 14 by __ pairs of selected array blocks in each memory frame. Preferably, the selected locations are 八孙八士,疋位Two blocks distributed in a rack, a block having the row decoders of the row decoders and the associated bit lines of the data sinks and the second blocks It is selected by another line and the greedy "flow bar", so that each rack doubles the bandwidth, but the current in any word, the film &amp; the moving current does not change. In addition, the 123178.doc in a selected row position -48- 200826114 The bit line - or many of them are also selected for resetting the stylization. At the same time, the number of processes may be limited by the current flowing from the selected bit lines in a block to the shared word line. However, this limitation is mitigated in a method in which some of the bits of the bit are reset to a higher resistance state, and the current through the 6 reset &quot;single 70 is reduced, along the IR of the shared word line segment. The drop is reduced and the remaining bits get more voltage to facilitate their reset. The preferred lines in each selected block (4) are all on the same column ^, thereby eliminating the decoding implication, since the global column decoder current does not need to be k-ized to support this point. Preferably, the simultaneously selected blocks are adjacent, in particular, in the case where word lines are shared between adjacent blocks. The decoding is configurable such that for any selected word line shared between two adjacent blocks, the two adjacent array blocks can be configured to simultaneously select the array block. For example, a given word line driver placed between the second and second blocks drives a common word line within the first and second blocks (both simultaneously selected). The next word line (assuming it is in the left and right sides of the array blocks in a 2: 丨 interleaved form) will be driven by an array of line drivers between the second and third array blocks, the array areas The block can also be selected as an array block. This avoids processing the selected word line from extending into the adjacent non-selected array block. When using the reset stylization, each memory unit is set back to a low resistance state by the "set", operation mode, which can be used to rewrite new data, or by applying a positive one time. To bias a bit to a bit, bite a number of bits in a data page or a erase block to erase a group of bits. Efficient erase can be selected by selecting multiple bit lines in a block Or a plurality of word lines, and set the special element to a low resistance. The current limiting circuit in the bit driver drive 123178.doc -49- 200826114 limits the total current flowing to the common word line. The selected memory cell technology, the relative number of set currents and reset currents, and the amount of u-cell leakage current can be used to set or erase operations by selecting fewer blocks for reset (ie, stylized). The selection of the resistive material is to form a diode polycrystalline germanium material. An antifuse ('AF'') can be serialized with the polycrystalline germanium diode, and the antifuse is in a stylizing event during a formatting step in manufacturing. Before the jump. This anti-fuse is used to limit the setting. The maximum current of the meta-conduction. As described above, preferably, the memory array includes a segmented word line architecture (as shown in Figures 12 and 13), and is preferably a three-dimensional array. In a particular embodiment, The word lines on a given word line layer are associated with bit lines on a single bit line layer, and in a particular embodiment, in a so-called "semi-mirror" configuration - the word lines on a given word line layer are shared between a two bit line layer (i.e., a single word line layer defining a two memory plane and a two bit line layer). Such a memory array structure further illustrates In the aforesaid U.S. Patent No. 6,879,505, the description of the various decoder circuits has been focused on a single array block. It should be recalled that each decoder has been in the _ source selection bus and for some embodiments - The background of the reverse source selection bus is described. The word line decoding hierarchy can be regarded as relatively direct. The source bus and the untwisted (four) line or the reverse miscellaneous selection confluence (4) are decoded based on the address information, and according to which array area Block system is valid _. Ben The other column decoding circuit has been referenced elsewhere. The (other) individual source selection bus and/or unselected bias line for the word line associated with the unshirted array block can be used. 123178.doc • 50- 200826114 Floating to the left. With respect to these row decoder configurations, a hierarchical bus configuration can be used to provide efficient routing of read/write data and effective biasing of bit lines within selected and unselected array blocks. A useful hierarchical bus configuration is illustrated in the context of the dual source select bus decoder shown in Figures 9 and 10, but these configurations are adjustable for use with other decoder specific embodiments. In the read and set, an exemplary hierarchical bus arrangement provides an appropriate bias on the SELN bus for a selected array block and causes the SELN bus of the unselected array block to float. This helps to reduce the undesirable power consumption in the array blocks within a selected one of the selected array blocks. The unselected word lines in a selected array block are biased at a relatively high voltage VUX (e.g., VPP_VT), and under a common word line architecture, the unselected word lines also extend to adjacent The selected array block (i.e., half of the word lines within the unselected array block are shared with the selected array block). The unselected locating elements within the adjacent array block are preferably biased at the unselected locating element line voltage VUB (e.g., ντ). This is a waste of power due to leakage current through unselected memory cells. The other half of the word lines in adjacent unselected array blocks float so that they leak until the VUB voltage, so the leakage power is minimized for half of the unselected cells. The exemplary hierarchical bus configuration also provides a vehicle and a long SELN.path&apos; that spans a number of blocks to reach the reset data drivers distributed under the array blocks. The four exemplary hierarchical bus arrangements are described in the next four figures. Referring now to Figure 19, a bus arrangement 500 is illustrated that includes three memory array blocks 502, 504, 506 that represent all of the array blocks within a rack. Although only three array blocks are shown, the incremental nature of the configuration and its extensibility to the number of any array block should be clear. The individual SELN bus segments for each of the array blocks are displayed. As used herein, a busbar segment is only a busbar that is smaller than other such busbars, while in other embodiments (described below), multiple busbar segments can be coupled together.

以形成單一更大匯流排。 在没定模式下,用於一選定陣列區塊之SELN匯流排片 段係耦a至一更長GSELN匯流排,其藉由一耦合電路5〇8 而跨越整個記憶體機架。此耦合電路5〇8可簡單至16個電 曰曰體各將一個別SELN匯流排線耦合至個別GSELN匯流 排線。此耦合電路508係由一控制信號EN—GSELN來致 旎,其在設定模式或在重置模式時有效用於選定陣列區塊 (下述)。在該設定模式期間,此GSELN匯流排係耦合至未 選定位元線電壓VUB(即該GSELN匯流排之各匯流排線係 耦合至此電壓)。用於該等未選定陣列區塊之個別 EN一GSELN控制信號係有效,個別耦合電路5〇8關閉,因 而需要時使個別SELN匯流排片段浮動。 在重置模式下’用於所有陣列區塊之個別En—GSELN控 制佗號係有效’且個別耦合電路5〇8係開啟以將個別seln 匯流排片段轉合至該GSELN匯流排。此點提供寫入資料至 所有陣列區塊’不管選擇哪個區塊。該SELB匯流排係驅 動至該VUX電壓(例如接地)以提供未選定位元線偏壓條件 123178.doc -52- 200826114 用以重置程式化。 此係一相對簡單的電路配置,每陣列區塊(耦合電路 508)僅需要一另外16個全域線(GSELN)及16個額外電晶 體。缺點(至少相對於下述其他具體實施例而言)包括在該 等SELB及SELN匯流排二者上的一相對較高電容。在該 SELB匯流排上的電容始終存在,但僅在一讀取循環期間 確定,而在所有該等SELN匯流排片段係耦合至全域匯流 排GSELN時,此時期間該等組合匯流排傳遞重置資料資 訊,在該SELN匯流排上的較高電容在該重置模式期間存 在。 在特定其他具體實施例中,該重置模式可組態有整個非 負電壓,而不將重置電壓VRR分割成- VRR/2及+VRR/2。 在此情況下,該等未選定字線及位元線係在中點(現在係 VRR/2)下偏壓。因此,當從重置模式出來時,應小心控制 該些線之放電速率以免在放電時過多電流突波。 現在參考圖20,描述另一具體實施例,其中該等個別 SELN匯流排片段係一起耦合以形成一單一更大匯流排, 其跨越整個記憶體機架。在設定模式下,用於一選定陣列 區塊之SELN匯流排係藉由一耦合電路532而耦合至一單一 偏壓線VUB,其跨越整個記憶體機架。此耦合電路532可 簡單至16個電晶體,各將一個別SELN匯流排線耦合至該 VUB偏壓線(其係耦合至一適當偏壓電路,如所示)。此耦 合電路532係由一控制信號BLATVUB來致能,其係在設定 模式時針對該選定陣列區塊有效。對於該等未選定陣列區 123178.doc •53- 200826114 塊,該個別BLATVUB控制信號係無效,個別耦合電路532 關閉’因而需要時使個別SELn匯流排片段浮動。 在重置模式下,該SELB匯流排係驅動至該VUB電壓(例 如接地)以提供未選定位元線偏壓條件用於重置程式化。 此外°亥專個別SELN匯流排片段係藉由一 |禺合電路μ3 一 起耦合以形成一跨越整個記憶體機架之單一匯流排,該記 憶體機架係耦合至該重置電路以向該等組合匯流排提供重 置資料負讯&quot;亥荨SELN匯流排片段之一可藉由匯流排53ό 而耦合至該重置電路。在特定具體實施例中,一耦合電路 535可用以在RESET模式下向重置區塊提供連接。 此係一相對簡單的電路配置,其每陣列區塊(該等耦合 電路532、533)僅需要一額外偏壓線(VUB)&amp;32個額外電晶 體。類似於先前具體實施例,在二SELB與SELN匯流排上 仍存在一相對較高電容。 現在參考圖21,描述一匯流排配置55〇,其併入來自二 先前具體實施例之特徵。在SET模式下,用於一選定陣列 區塊之SELN匯流排片段係藉由一耦合電路554而耦合至一 跨越整個記憶體機架之VUB偏壓線,耦合電路554係由一 控制信號BLATVUB來㈣。用於未選定陣歹,】區塊之個別 BLATVUB控制信號係無效,個別耦合電路554關閉,因而 需要時使個別SELN匯流排片段浮動(由於在SET模式下該 EN一GSELN信號也無效)。 在重置模式下’用於一選定陣列區塊之個別抓―仍腳 控制信號係有效,故一個別耦合電路552係開啟以將個別 123178.doc -54- 200826114 SELN匯流排片段耦合至該GSELN匯流排。用於該等未選 定陣列區塊之個別EN—GSELN控制信號係無效,個別耦合 電路552關閉,故使個別SELN匯流排片段浮動。此組態僅 向該(等)未選定陣列區塊提供寫入資料,從而明顯減小總 電容。該SELB匯流排係驅動至該VUX電壓(例如接地)以提 供未選定位元線偏壓條件用於重置程式化。 此電路配置每陣列區塊(該等耦合電路552、554)需要17 個額外線(VUB匯流排與GSELN匯流排)與32個額外電晶 體。不同於該等先前具體實施例,此配置提供用於明顯減 小SELN匯流排上的電容,由於用於未選定陣列區塊之該 等個別SELN匯流排片段不耦合至該GSELN匯流排。在該 SELB匯流排上仍存在相當南的電容。 圖22描述另一階層式匯流排配置,此次僅利用一跨越記 憶體機架的單一全域選擇匯流排GSEL,並將SELB匯流排 分成一個別SELB匯流排片段用於各陣列區塊。對於一選 定陣列區塊,個別SELB匯流排或個別SELN匯流排片段係 耦合至此GSEL匯流排。在SET模式期間,該選定區塊 SELB匯流排片段係耗合至該GSEL匯流排,而該選定區塊 SELN匯流排片段係耦合至該VDSEL·偏壓線(在SET期間其 傳遞一適當偏壓電路所產生之未選定位元線偏壓條件 VUB,如所示)。使該等未選定區塊SELN匯流排向左浮 動。 在RESET模式期間,該選定區塊SELN匯流排片段係耦 合至該GSEL匯流排,而該選定區塊SELB匯流排片段係耦 123178.doc -55- 200826114 合至該VDSEL偏壓線(在rESET期間其傳遞未選定位元線 偏壓條件νυχ)。再次使該等未選定區塊SELN匯流排向左 浮動。 此配置係所述该等配置中最複雜的,每陣列區塊需要17 個全域線(即跨越記憶體機架)與64個額外電晶體,且可能 在某些具體實施例中需要更多佈局面積。然而,其還在 SELB及SELN匯流排上提供低電容,因為允許更高的效 能,並提供一極模組化的區塊設計。而且,可實施更大記 憶體機架而不明顯增加該等SELB及SELN匯流排上的電 容。 在另一具體實施例中,該等行解碼器電路可加以修改以 提供一分離行解碼輸出用於位元線驅動器電路之該等 NMOS及PMOS電晶體,故可將位元線選擇器設定在高阻 抗狀悲下。但此配置將明顯增加位元線選擇器面積以及自 身的行解碼器。 現在參考圖23,描述一資料電路,其包括用於該等設 疋、重置及讀取模式之分離區塊。應記得,在反向偏壓模 式(即重置模式),該等選定位元線係耦合至一個別SELN匯 流排線(即該反向源極選擇匯流排)。此處發現一重置驅動 器61 5耦合至SELN匯流排617(其表示可使用四個階層式匯 流排配置之任一者的至冗!^^匯流排之路徑)。本質上,此 表不對於一選定陣列區塊最終耦合至seln匯流排片段之 路裣。奴寫入資料資訊係接收於&quot;〇邏輯6〇【,在匯流排 602上傳遞至一寫入鎖存器區塊6〇4,在匯流排上傳遞 123178.doc -56- 200826114 至控制邏輯60 8 ’該控制邏輯接著藉助控制線6丨2來控制重 置驅動器61 5。 應記得’在該正向模式下,該等選定位元線係耦合至一 個別SELB匯流排線。由於二SEt及READ模式利用正向偏 壓模式,故一設定驅動器614及一讀取感應放大器613係同 時柄合至SELB匯流排616(其表示用於上述四個階層式匯流 排配置之任一者或可採用之任何其他配置之至SELB匯流 排之路徑)。感應資料係由匯流排609而傳遞至一讀取鎖存 器605,由匯流排603而傳遞至I/O邏輯601。各種匯流排 606、6 10及611提供用於一程式化控制迴路,有時稱為智 慧寫入’其可在一位元係成功跳變或設定時關閉程式化電 流。該等匯流排還提供一寫入前讀取能力以決定(例如)在 後續程式化操作之前應保留之任何先前程式化狀態(例 如LSB資料位元)。此類能力係進一步說明於下面參考的 023-0049及 023-0055 申請案中。 圖24中描述一簡化範例性重置驅動器6丨5以及至一選定 記憶體單元638之字線及位元線選擇路徑之一表示。一字 線選擇路徑639表示透過該字線驅動器電路(即解碼器頭)至 用於產生解碼源極選擇匯流排XSELN之電路的路徑。一位 元線選擇路徑636表示透過該位元線驅動器電路以及透過 任何搞合電路(例如在各種階層式匯流排配置具體實施例 中所述之該等電路)至個別SELN匯流排線635之路徑。一 較佳重置方法及相關聯重置驅動器係說明於下面參考的 SAND-01U4US0及SAND-01114US1申請案中,特別涉及其 123178.doc •57- 200826114 中圖13。 位元線選擇路徑之電容係在試圖程式化一新定址選定位 元線之前預充電。此點可使用一高於實際重置選定記憶體 單元所需之數量的電流來執行,但適當定時時,此類更高 數量預充電可加快預充電時間而對記憶體單元沒有決定性 的影響。此預充電係由一在控制信號637上傳遞至位元線 選擇路徑636之預充電行信號pCHGCOL來控制。一位元線 預充電(BLP)限流電路633與一重置限制電路634係同時提 供以控制個別位元線預充電及重置電流之上部數量。若資 料使得不需要任何重置操作,則二者均由信號632來停 用,且SELN匯流排線635浮動。 反之’若資料使得要重置記憶體單元,則停用線632係 無效’且暫時(例如200 ns至5〇〇 ns)致能BLP限流電路633 以提供一更高位準的控制電流用於此類預充電,之後其係 停用(藉由一未顯示控制信號),使重置限流電路634來供應 一更低數量的電流用於重置選定記憶體單元。由於重置一 記憶體單元引起其從一更低電阻狀態變成一更高電阻狀 態,故很少需要感應重置操作之完成並停用重置限制 634,由於單元在其到達重置狀態便自動關閉。 至於上述各種具體實施例,許多類型記憶體單元能夠使 用一反向偏壓(例如上述重置模式)來加以程式化。此類單 元包括一被動元件單元,其具有一金屬氧化物(例如一過 渡金屬氧化物)與一二極體。其他適當單元包括在一二極 體矩陣内的一電阻材料之該等單元。範例包括一可程式化 123178.doc -58- 200826114 金屬化連接、一相變電阻器(例如GST材料)、一有機材料 可變電阻器、一複合金屬氧化物、一碳聚合物膜、一播雜 硫化物玻璃及一含遷移原子以改變電阻之宵特基 (Schottky)阻障二極體。所選電阻材料可提供一次可程^ 化(0TP)記憶體單元或多次寫入記憶體單元。此外,可採 用一多晶矽二極體,其具有反向偏壓應力修改的傳導。 ‘用於反向重置操作之有用記憶體單元係說明於授予s.To form a single larger bus. In the undefined mode, the SELN busbar segments for a selected array block are coupled to a longer GSELN busbar that spans the entire memory bay by a coupling circuit 5〇8. The coupling circuit 5〇8 can be as simple as 16 electrons each coupling a different SELN bus line to an individual GSELN bus line. This coupling circuit 508 is activated by a control signal EN_GSELN which is effective for the selected array block (described below) in the set mode or in the reset mode. During this set mode, the GSELN busbar is coupled to the unselected location line voltage VUB (i.e., the busbar lines of the GSELN bus are coupled to this voltage). The individual EN-GSELN control signals for the unselected array blocks are active, and the individual coupling circuits 5〇8 are turned off, thereby allowing individual SELN bus-segment segments to float as needed. In the reset mode, the individual En-GSELN control registers for all of the array blocks are active and the individual coupling circuits 5〇8 are turned on to switch the individual selln bus segments to the GSELN bus. This point provides write data to all array blocks' regardless of which block is selected. The SELB bus is driven to the VUX voltage (e.g., ground) to provide unselected positioning element bias conditions 123178.doc -52 - 200826114 to reset the stylization. This is a relatively simple circuit configuration, requiring only an additional 16 global lines (GSELN) and 16 additional transistors per array block (coupling circuit 508). Disadvantages (at least with respect to other embodiments described below) include a relatively high capacitance on both the SELB and SELN busbars. The capacitance on the SELB bus is always present, but only during a read cycle, and when all of the SELN bus segments are coupled to the global bus GSELN, the combined bus transfers are reset during this time. Data information, the higher capacitance on the SELN bus is present during this reset mode. In certain other embodiments, the reset mode can be configured with the entire non-negative voltage without dividing the reset voltage VRR into -VRR/2 and +VRR/2. In this case, the unselected word lines and bit lines are biased at the midpoint (now VRR/2). Therefore, when coming out of the reset mode, the discharge rate of the lines should be carefully controlled to avoid excessive current surges during discharge. Referring now to Figure 20, another embodiment is described in which the individual SELN busbar segments are coupled together to form a single larger busbar that spans the entire memory bay. In the set mode, the SELN bus for a selected array block is coupled to a single bias line VUB by a coupling circuit 532 that spans the entire memory frame. The coupling circuit 532 can be as simple as 16 transistors, each coupling a different SELN bus bar to the VUB bias line (which is coupled to a suitable bias circuit, as shown). The coupling circuit 532 is enabled by a control signal BLATVUB that is active for the selected array block when in the set mode. For the unselected array regions 123178.doc • 53- 200826114, the individual BLATVUB control signals are disabled and the individual coupling circuits 532 are turned off&apos; thus floating individual SELn bus segments as needed. In the reset mode, the SELB bus is driven to the VUB voltage (e.g., ground) to provide an unselected positioning element bias condition for resetting the stylization. In addition, the individual SELEN busbar segments are coupled together by a coupling circuit μ3 to form a single busbar spanning the entire memory frame, the memory chassis being coupled to the reset circuit for The combined bus provides a reset data negative &quot; one of the HELN SELN bus segments can be coupled to the reset circuit by a bus bar 53A. In a particular embodiment, a coupling circuit 535 can be used to provide a connection to the reset block in RESET mode. This is a relatively simple circuit configuration in which only one additional bias line (VUB) &amp; 32 additional transistors are required per array block (the coupling circuits 532, 533). Similar to the previous embodiment, there is still a relatively high capacitance on the two SELB and SELN busbars. Referring now to Figure 21, a busbar configuration 55A is described which incorporates features from two prior embodiments. In SET mode, the SELN bus segment for a selected array block is coupled to a VUB bias line across the entire memory frame by a coupling circuit 554, which is coupled by a control signal BLATVUB. (4). For unselected arrays, the individual BLATVUB control signals are not valid, and the individual coupling circuits 554 are turned off, thus allowing individual SELN bus segments to float as needed (since the EN-GSELN signal is also inactive in SET mode). In the reset mode, the 'single-slide control signal for a selected array block is valid, so a separate coupling circuit 552 is turned on to couple the individual 123178.doc -54 - 200826114 SELN bus segment to the GSELN. Bus bar. The individual EN-GSELN control signals for the unselected array blocks are invalid, and the individual coupling circuits 552 are turned off, thereby causing individual SELN bus segments to float. This configuration only provides write data to this (etc.) unselected array block, which significantly reduces the total capacitance. The SELB bus is driven to the VUX voltage (e. g., ground) to provide unselected positioning element bias conditions for reset stylization. This circuit configuration requires 17 additional lines (VUB bus and GSELN bus) and 32 additional transistors for each array block (the coupling circuits 552, 554). Unlike these prior embodiments, this configuration provides for significantly reducing the capacitance on the SELN bus, since the individual SELN bus segments for unselected array blocks are not coupled to the GSELN bus. There is still a fairly south capacitance on the SELB bus. Figure 22 depicts another hierarchical bus arrangement, this time using only a single global selection bus GSEL across the memory frame and dividing the SELB bus into a separate SELB bus segment for each array block. For a selected array block, individual SELB busses or individual SELN bus segments are coupled to the GSEL bus. During the SET mode, the selected block SELB busbar segment is consuming to the GSEL busbar, and the selected block SELN busbar segment is coupled to the VDSEL. bias line (which transmits an appropriate bias during SET) The unselected positioning element line bias condition VUB generated by the circuit, as shown). The unselected block SELN bus is floated to the left. During the RESET mode, the selected block SELN bus segment is coupled to the GSEL bus, and the selected block SELB bus segment is coupled to the VDSEL bias line (during rESET) It passes the unselected positioning element line bias condition νυχ). The unselected block SELN bus is again floated to the left. This configuration is the most complex of the described configurations, requiring 17 global lines per array block (ie, spanning the memory rack) and 64 additional transistors, and may require more layout in some embodiments. area. However, it also provides low capacitance on the SELB and SELN busbars because it allows for higher efficiency and provides a one-pole modular block design. Moreover, larger memory racks can be implemented without significantly increasing the capacitance on the SELB and SELN busbars. In another embodiment, the row decoder circuits can be modified to provide a separate row decode output for the NMOS and PMOS transistors of the bit line driver circuit, so that the bit line selector can be set at High impedance and sadness. However, this configuration will significantly increase the bit line selector area as well as its own row decoder. Referring now to Figure 23, a data circuit is illustrated that includes separate blocks for the set, reset, and read modes. It will be recalled that in the reverse bias mode (i.e., reset mode), the selected positioning elements are coupled to a different SELN bus line (i.e., the reverse source select bus). Here, it is found that a reset driver 61 5 is coupled to the SELN bus 617 (which indicates that the path to any of the four hierarchical bus configurations can be used). Essentially, this table does not ultimately couple to the Sarn bus segment for a selected array block. The slave write data information is received in the &quot;〇 logic6〇[, on the bus 602 is passed to a write latch block 6〇4, and 123178.doc -56-200826114 is passed on the bus to the control logic 60 8 'The control logic then controls the reset driver 61 5 by means of the control line 6丨2. It should be remembered that in this forward mode, the selected positioning elements are coupled to an individual SELB bus line. Since the two SEt and READ modes utilize the forward bias mode, a set driver 614 and a read sense amplifier 613 are simultaneously spliced to the SELB bus bar 616 (which represents any of the above four hierarchical bus bar configurations). Or any other configuration to the path of the SELB bus). The sensing data is passed from bus 609 to a read latch 605 for transfer to I/O logic 601 by bus 603. Various bus bars 606, 6 10 and 611 are provided for a stylized control loop, sometimes referred to as a smart write, which can turn off the stylized current when a meta-system is successfully toggled or set. The bus banks also provide a pre-write read capability to determine, for example, any previous stylized states (e.g., LSB data bits) that should be retained prior to subsequent stylization operations. Such capabilities are further described in the 023-0049 and 023-0055 applications referenced below. A simplified exemplary reset driver 6丨5 and a word line and bit line select path representation of a selected memory cell 638 are depicted in FIG. A word line select path 639 represents the path through the word line driver circuit (i.e., the decoder head) to the circuitry used to generate the decoded source select bus XSELN. A one-element selection path 636 represents the path through the bit line driver circuit and through any of the matching circuits (e.g., the circuits described in the various hierarchical bus arrangement embodiments) to the individual SELN bus bars 635. . A preferred reset method and associated reset driver are described in the SAND-01U4US0 and SAND-01114US1 applications referenced below, in particular to Figure 13 of 123178.doc • 57-200826114. The capacitance of the bit line selection path is precharged before attempting to program a new address selection location line. This can be performed using a higher current than the amount required to actually reset the selected memory cell, but at the appropriate timing, such higher amounts of pre-charging can speed up the pre-charge time without a decisive influence on the memory cells. This precharge is controlled by a precharge line signal pCHGCOL that is passed on control signal 637 to bit line select path 636. A one-line precharge (BLP) current limit circuit 633 and a reset limit circuit 634 are simultaneously provided to control the number of upper bits of the individual bit line precharge and reset current. If the data is such that no reset operation is required, both are disabled by signal 632 and the SELN bus bar 635 is floating. Conversely, 'If the data causes the memory cell to be reset, the disable line 632 is disabled' and temporarily (eg, 200 ns to 5 ns) enables the BLP current limiting circuit 633 to provide a higher level of control current for Such pre-charging, after which it is disabled (by a control signal not shown), causes reset current limiting circuit 634 to supply a lower amount of current for resetting the selected memory cell. Since resetting a memory cell causes it to change from a lower resistance state to a higher resistance state, it is less necessary to perform the inductive reset operation and disable the reset limit 634 since the cell automatically reaches its reset state. shut down. As with the various embodiments described above, many types of memory cells can be programmed using a reverse bias (e.g., the reset mode described above). Such a unit includes a passive component unit having a metal oxide (e.g., a transition metal oxide) and a diode. Other suitable units include such elements of a resistive material within a matrix of diodes. Examples include a programmable 123178.doc -58- 200826114 metallization connection, a phase change resistor (such as GST material), an organic material variable resistor, a composite metal oxide, a carbon polymer film, a broadcast Heterosulfide glass and a Schottky barrier diode containing a migrating atom to change the electrical resistance. The selected resistive material provides one-time programmable (0TP) memory cells or multiple writes to the memory cells. In addition, a polysilicon diode can be used which has a reverse bias stress modified conduction. ‘The useful memory unit for reverse reset operation is described in grant s.

Brad Herner等人之美國專利第6,952,〇3〇號,標題為,,高密 度二維記憶體單元”;以及還說明於2〇〇5年12月Μ日 ; Tanmay Kumar等人申請的美國申請案第11/237,167號,標 . 題為’,使用帶可微調電阻之可切換半導體記憶體元件之記 憶體單元之方法’,,於2007年4月26曰作為美國專利申請公 告案第2007-0090425號公佈。一適當金屬氧化物記憶體單 元係顯示於2006年3月31曰S. Brad Herner申請的美國申請 案苐11/394,903號,標題為”含電阻率切換氧化物或氮化物 及反絲之多層非揮發性記憶體單元”。一使用一可提供 多個電阻狀態之相變材料之適當記憶體單元係顯示於R〇y Ε· Scheuerlein等人申請的美國專利申請公告案第2〇〇5_ 015 8 9 5 0號’標題為π串列包含介電層及相變材料之非揮發 - 性圮憶體單元’’。該些上述參考揭示案之各案全體内容以 引用形式併入本文。具有一過渡金屬氧化物(例如包括該 等具有鈷之氧化物)之其他範例性記憶體單元以及其中操 縱元件之多晶矽材料自身包含可切換電阻材料之範例性單 元係說明於下面所參考之ΜΑ-163-1申請案中。 123178.doc •59- 200826114 此外,S. Brad Herner等人於2005年5月9曰申請的美國申 請案第11/125,939號,標題為”包含二極體及電阻切換材料 之可再寫記憶體單元,,,於2006年12月9日作為美國專利申 請公告案第2006-0250836號公佈,揭示一串列一氧化物(氧 化鎳)併入一二極體之有用可再寫記憶體單元,其中該記 憶體單元之電阻可從低至高及從高至低電阻狀態而重複切 換。S. Brad Herner等人於20 06年3月31申請的美國申請案 第11/395,995號,標題為”包含二極體及電阻切換材料之非 揮發性記憶體單元”並於2006年11月9日作為美國專利申請 公告案第2006-0250837號而公佈,揭示一 οτρ多層記憶體 單元’其係使用正向偏壓來設定並使用反向偏壓來重置。 該些上述參考揭示案之各案全體内容以引用形式併入本 文。 在本文所述許多具體實施例中,在資料路徑内強加於各 個別匯流排線之精確偏壓條件係獨立控制。用於該等設定 及重置驅動器之各驅動器的特定電壓及電流設定可針對資 料路徑之各位元來加以調整。由此,涵蓋具有兩個以上狀 態之特定記憶體單元(即”多層”記憶體單元)用於配合本文 所述許多結構來使用。範例性多層記憶體單元係說明於下 面參考的前述美國申請案第申請 案中。 可用於實施本發明之範例性被動元件記憶體單元及相關 非揮發性記憶體結構係說明下列文件中,其全部内容各以 引用形式併入本文: 123178.doc -60- 200826114 美國專利第6,034,882號,標題為&quot;垂直堆疊場可程式化 非揮發性記憶體及製造方法”,授予歸G jQhns〇n等 人; 美國專利第6,420,215號,標題為”三維記憶體陣列及製 造方法&quot;,授予N· Johan Knall等人; 美國專利第6,525,953號,標題為&quot;垂直堆疊場可程式化 非揮發性記憶體及製造方法&quot;,授予MarkJ〇hns〇n等人; 美國專利第6,490,218號,標題為&quot;用於儲存多位元數位 資料之蘇會為記憶體方法及系統”,授予Michaei Vyv〇da 等人; 美國專利第6,952,043號,標題為,,主動裝置中的電絕緣 柱” ’授予Michael Vyvoda等人;以及 美國專利申請公告案第1182〇〇5-〇〇52915號,標題為,,具 有高及第阻抗態之不帶介電反溶絲之非揮發性記憶體單 元’’,由S· Brad Herner等人申請。 下列申請案(各在相同時期申請)說明可用於實施本發 明之記憶體單元結構、電路、系統及方法,各申請案全 部内容以引用形式併入本文: 美國申請案第11/496,985號(律師檔第10519-141號),標 題為”多用途記憶體單元及記憶體陣列’’,由Roy Scheuerlein與 Tanmay Kumar 申請(&quot;10519_141” 申請案); 美國申請案第11/496,984號(律師檔案號105 19-150),標 題為”多用途記憶體單元及記憶體陣列之使用方法”,由 Roy Scheuerlein與 Tanmay Kumar 申請(’’10519-150” 申請 123178.doc -61 - 200826114 案); 美國申請案第11/496,874號(律師檔案號10519-142),標 題為’’混合用途記憶體單元π,由R〇y Scheuerlein申請 (’’10519-142” 申請案); 美國申請案第11/496,983號(律師檔案號10519-151),標 題為”混合用途記憶體單元之使用方法”,由R〇yU.S. Patent No. 6,952, 〇 3, issued to Brad Herner et al., entitled "High Density Two-Dimensional Memory Unit"; and also on December 30, 2005; US application by Tanmay Kumar et al. Case No. 11/237,167, entitled "Method of Using Memory Units with Switchable Semiconductor Memory Elements with Trimmer Resistors", April 26, 2007 as US Patent Application Bulletin No. 2007-0090425. A suitable metal oxide memory cell system is shown in U.S. Application Serial No. 11/394,903, filed on March 31, 2006 by S. Brad Herner, entitled "Resistivity Switching Oxide or Nitride" And a multi-layer non-volatile memory cell of the anti-filament." A suitable memory cell system using a phase change material that provides a plurality of resistive states is shown in U.S. Patent Application Publication No. 2〇〇5_ 015 8 9 5 0 'The title is π tandem containing the dielectric layer and the non-volatile memory of the phase change material'. The entire contents of the above referenced disclosures are by reference. Incorporated into this article. Has one Other exemplary memory cells that pass metal oxides (eg, including such oxides with cobalt) and exemplary cells in which the polycrystalline germanium material of the steering element itself comprises a switchable resistive material are described in the above-referenced ΜΑ-163- 1 Application. 123178.doc •59-200826114 In addition, US Application No. 11/125,939, filed May 29, 2005, to S. Brad Herner et al., entitled "Including Diodes and Resistor Switching The rewritable memory unit of the material is disclosed in U.S. Patent Application Publication No. 2006-0250836, issued on Dec. 9, 2006, which discloses the utility of a tantalum oxide (nickel oxide) incorporated into a diode. The memory unit can be rewritten, wherein the resistance of the memory unit can be repeatedly switched from low to high and from high to low resistance. U.S. Application Serial No. 11/395,995, filed on Mar. 31, 2006, entitled,,,,,,,,,,,,,,,,,,, Japanese Patent Application Publication No. 2006-0250837 discloses a οτρ multilayer memory unit that is set using a forward bias and reset using a reverse bias. The entire contents of each of the above-referenced publications are hereby incorporated by reference. In many of the embodiments described herein, the precise bias conditions imposed on each individual bus bar within the data path are independently controlled. The specific voltage and current settings for each of these drivers for setting and resetting the drive can be adjusted for each element of the data path. Thus, a particular memory unit (i.e., "multi-layer" memory unit) having more than two states is contemplated for use with the many structures described herein. Exemplary multilayer memory cells are described in the aforementioned U.S. Application Serial No. Exemplary passive element memory cells and associated non-volatile memory structures that can be used in the practice of the present invention are described in the following documents, the entire contents of each of which are hereby incorporated by reference: s. , titled &quot;Vertical stacking field, programmable non-volatile memory and manufacturing method," awarded GjQhns〇n et al; US Patent No. 6,420,215, entitled "Three-Dimensional Memory Array and Manufacturing Method", awarded N. Johan Knall et al.; U.S. Patent No. 6,525,953, entitled &quot;Vertical Stacking Field Programmable Non-Volatile Memory and Manufacturing Method&quot;, awarded to Mark J〇hns〇n et al; US Patent No. 6,490,218, title For &quot;Su will be a memory method and system for storing multi-bit digital data," issued by Michaei Vyv〇da et al; US Patent No. 6,952,043, entitled "Electrical Insulation Columns in Active Devices" Michael Vyvoda et al; and U.S. Patent Application Bulletin No. 1182〇〇5-〇〇52915, entitled, with high and first impedance states Electric anti-solvent, filament non-volatile memory unit ', filed by S · Brad Herner et al. The following applications (each filed during the same period) describe the memory cell structures, circuits, systems, and methods that can be used to implement the present invention, and the entire contents of each application are hereby incorporated by reference: U.S. Application No. 11/496,985 (Attorney) File No. 10519-141, entitled "Multi-Purpose Memory Units and Memory Arrays", by Roy Scheuerlein and Tanmay Kumar (&quot;10519_141" Application); US Application No. 11/496,984 (Attorney Profile) No. 105 19-150), entitled "Usage of Multi-Purpose Memory Units and Memory Arrays", by Roy Scheuerlein and Tanmay Kumar (''10519-150' application 123178.doc -61 - 200826114); Application No. 11/496,874 (Attorney Docket No. 10519-142), entitled ''Mixed-use memory unit π, applied by R〇y Scheuerlein (''10519-142' application); US Application No. 11/ No. 496,983 (Attorney Docket No. 10519-151), entitled "How to Use Mixed-Use Memory Units", by R〇y

Scheuerlein申請(&quot;10519-15Γ 申請案); 美國申請案第11/496,870號(律師檔案號1〇519_149),標 題為”具不同資料狀態之混合用途記憶體單元,,,由R〇yApplication by Scheuerlein (&quot;10519-15Γ application); US Application No. 11/496,870 (Attorney Docket No. 1〇519_149), entitled "Mixed-use memory unit with different data status,, by R〇 y

Scheuerlein 與 Christopher Petti 申請(&quot;10519-149” 申請 案); 美國申請案第11/497,021號(律師檔案號10519_152),標 題為”具不同資料狀態之混合用途記憶體單元之使用方 法”,由 Roy Scheuerlein 與 Christopher Petti 申請(π 105 19- 152”申請案); 美國申請案第11/461,393號(律師檔案號SAND-01114US0), 標題為”在非揮發性記憶體中的受控脈衝操作,,,由R〇yScheuerlein and Christopher Petti Application (&quot;10519-149" Application); US Application No. 11/497,021 (Attorney Docket No. 10519_152), entitled "Usage Method for Mixed-Use Memory Units with Different Data Status", by Roy Scheuerlein and Christopher Petti Application (π 105 19-152); US Application No. 11/461,393 (Attorney Docket No. SAND-01114US0), entitled "Controlled Pulse Operation in Non-Volatile Memory, , by R〇y

Scheuerlein 申請(&quot;SAND-01114US0’,申請案); 美國申請案第1 1/461,399號(律師檔案號8八通_ 01114US 1 ),標題為”用於非揮發性記憶體中受控脈衝操 作之糸統,由Roy Scheuerlein 申請(&quot;SAND-01114US1,, 申請案); 美國申請案第11/461,410號(律師檔案號sand· 01115US0),標題為&quot;高帶寬一次場可程式化記憶體,,, 123178.doc -62- 200826114 由 Roy Scheuerlein 與 Christopher J· Petti 申請(’’SAND-01115 U S 0’’申請案 ) ; 美國申請案第11/461,419號(律師檔案號SAND-0111 5US 1),標題為”用於高帶寬一次場可程式化記憶體 之系統’’,由 Roy Scheuerlein與 Christopher J· Petti 申請 (’’SAND-0111 5USln申請案); 美國申請案第11/461,424號(律師檔案號SAND-0111 7US0),標題為”在非揮發性記憶體中的反向偏壓微 調操作’’,由 Roy Scheuerlein 與 Tanmay Kumar 申請 (’’SAND-01117US0”申請案); 美國申請案第11/461,431號(律師檔案號SAND-01117US 1),標題為”用於非揮發性記憶體中反向偏壓微 調操作之系統&quot;,由Roy Scheuerlein與Tanmay Kumar申 請(ΠSAND-01117USΓ’申請案); 美國申請案第11/496,986號(律師檔案號厘八-163-1),標 題為’’包含具可微調電阻之可切換半導體記憶體元件之 記憶體單元之使用方法·’,由Tanmay Kumar、S. Brad Herner、Roy E. Scheuerlein及 Christopher J. Petti 申請 (&quot;MA-163-1”申請案); 美國申請案第11/461,339號(律師檔案號023-0048),標題 為’’併入反向極性字線及位元線解碼器之被動元件記憶 體陣列π,由Luca G. Fasoli、Christopher J· Petti 及 Roy Ε· Scheuerlein 申請(π〇23-0048π 申請案); 美國申請案第11/461,364號(律師檔案號023-0054),標題 123178.doc -63 - 200826114 為π使用併入反向極性字線及位元線解碼器之被動元件 各己憶體陣列之方法’’,由Luca G. Fasoli、Christopher J. Petti及 Roy Ε· Scheuerlein 申請(’’023-0054” 申請案); 美國申請案第11/461,343號(律師檔案號023-0049),標題 為π用於讀取一多層被動元件記憶體單元陣列之裝置,’, 由 Roy Ε. Scheuerlein、Tyler Thorp及 Luca G· Fasoli 申請 (&quot;023-0049” 申請案); 美國申請案第11/461,367號(律師檔案號023-0055),標題 為’’用於讀取一多層被動元件記憶體單元陣列之方法”, 由 Roy E· Scheuerlein、Tyler Thorp及 Luca G. Fasoli 申請 (&quot;023-0055”申請案); 美國申請案第11/461,352號(律師檔案號023-0051),標題 為π用於耦合讀取/寫入電路至記憶體陣列之雙資料相依 匯流排’’,由 Roy Ε. Scheuerlein 及Luca G· Fasoli 申請 (&quot;023-0051”申請案); 美國申請案第11/461,369號(律師檔案號〇23-0056),標題 為π用於耦合讀取/寫入電路至記憶體陣列之雙資料相依 匯流排之使用方法’’,由Roy Ε. Scheuerlein及Luca G.Scheuerlein Application (&quot;SAND-01114US0', Application); US Application No. 1 1/461,399 (Attorney File No. 8 Octopus _ 01114US 1), entitled "Controlled Pulse Operation in Non-Volatile Memory糸, applied by Roy Scheuerlein (&quot;SAND-01114US1,, application); US application No. 11/461,410 (lawyer file number sand 01115US0), titled &quot;High bandwidth first field programmable Memory,,, 123178.doc -62- 200826114 Application by Roy Scheuerlein and Christopher J. Petti (''SAND-01115 US 0'' application); US application No. 11/461,419 (lawyer file number SAND-0111) 5US 1), entitled "System for High-Bandwidth One-Time Programmable Memory", by Roy Scheuerlein and Christopher J. Petti ('SAND-0111 5USln Application); US Application 11/461 , No. 424 (attorney file number SAND-0111 7US0), entitled "Reverse Bias Fine-Tuning Operation in Non-Volatile Memory", by Roy Scheuerlein and Tanmay Kumar (''SAND-01117US0') ; U.S. Application No. 11/461,431 (Attorney Docket No. SAND-01117US 1) entitled "System for Reverse Bias Fine-Tuning Operation in Non-Volatile Memory", applied by Roy Scheuerlein and Tanmay Kumar ( ΠSAND-01117USΓ' application; US Application No. 11/496,986 (Attorney Archive No. VIII-163-1), entitled "Use of Memory Units Containing Switchable Semiconductor Memory Elements with Trimmer Resistors" Method · ', applied by Tanmay Kumar, S. Brad Herner, Roy E. Scheuerlein and Christopher J. Petti (&quot;MA-163-1" application); US application No. 11/461,339 (lawyer file number) 023-0048), titled ''passive element memory array π incorporating reverse polarity word line and bit line decoder, applied by Luca G. Fasoli, Christopher J. Petti and Roy Ε Scheuerlein (π〇23) -0048π Application); US Application No. 11/461,364 (Attorney Docket No. 023-0054), Title 123178.doc -63 - 200826114 Incorporating Reverse Polarity Word Lines and Bit Line Decoders for π Use Method for passive array elements ', applied by Luca G. Fasoli, Christopher J. Petti and Roy Ε Scheuerlein (''023-0054' application)); US application No. 11/461, 343 (lawyer file number 023-0049), titled π is used to read a multi-layer passive element memory cell array, ', applied by Roy Ε. Scheuerlein, Tyler Thorp and Luca G. Fasoli (&quot;023-0049) application); US application 11/ 461,367 (attorney file number 023-0055), entitled ''Method for reading a multi-layer passive element memory cell array', by Roy E. Scheuerlein, Tyler Thorp and Luca G. Fasoli (&quot ; 023-0055" application; US Application No. 11/461,352 (Attorney Docket No. 023-0051), titled π for a dual data dependent bus for coupling read/write circuits to a memory array '', applied by Roy Ε. Scheuerlein and Luca G. Fasoli (&quot;023-0051" application); US application No. 11/461,369 (lawyer file number -023-0056), titled π for coupled reading Double data sharing of the fetch/write circuit to the memory array The use of row '', Ε the Roy. Scheuerlein and Luca G.

Fasoli 申請(”〇23-0056” 申請案); 美國申請案第11/461,359號(律師檔案號023-0052),標題 為”併入用於記憶體陣列區塊選擇之二資料匯流排之記 憶體陣列’’,由 Roy E. Scheuerlein、Luca G. Fasoli 及 Christopher J. Petti 申請(’’023-0052’’申請案); 美國申請案第11/46 1,372號(律師檔案號023-0057),標題 123178.doc -64- 200826114 為π用於記憶體陣列區塊選擇之二資料匯流排之使用方 法”,由 Roy Ε· Scheuerlein、Luca G. Fasoli 及 Christopher J· Petti 申請(”〇23-0057” 申請案); 美國申請案第11/46 1,362號(律師檔案號023-0053),標題 為π用於區塊可選擇記憶體陣列之階層式位元線偏壓匯 流排’’,由 Roy Ε. Scheuerlein及 Luca G· Fasoli 申請(,,〇23- 0053”申請案);以及 美國申請案第11/461,376號(律師檔案號023 — 0058),標題 為&quot;使用用於區塊可選擇記憶體陣列之階層式位元線偏 壓匯流排之方法&quot;,由Roy E. Scheuerlein及Luca G. Fasoli 申請(&quot;023-0058&quot;申請案)。 應瞭解,本文所示特定範例性具體實施例一直在特定數 位範例之背景下說明,例如解碼輸出之數目、解碼器頭之 數目、匯流排線之數目、資料匯流排之數目、在一記憶體 機架内陣列區塊之數目及記憶體條之數目。可使用此揭示 案之教導來實施符合其他設計目標之其他變化。清楚起 見,並未顯示並說明|文所述實&amp;方案之全部常規特徵。 大多數記憶體陣列係設計具有一相對較高的均勻度。例 如i通常每一位元線包括相同數目的記憶體單元。作為另 -乾例,纟元線、字線、陣列區塊及甚至記憶體平面之數 目上常係2的-整數次冪(例如’ 2n),以獲得解碼電路之簡 化,效率。但此類規則性或一致性對於本發明之任一具體 實知例中當然不要求。例如,在 — 括不同數目的記憶體單元,該記憶體陣列可包括三個記憶 123178.doc -65 - 200826114 體平面,在第一及最後陣列區塊内的字線片段可能在記憶 體單元數目或位元線組態及對記憶體陣列設計之通常一致 性的許多其他不規則變化之任一變化上不同,除非申請專 利範圍中另有明確說明,即便如本文所述具體實施例中所 示,此類通常規則性不應引入任何申請專利獨立項之意義 内。 應瞭解,指示頂部、左邊、底部及右邊僅係用於一記憶 體之四側之方便說明性術語。用於一區塊之該等字線片段 可實施為水平定向之二指間字線片段群組,而用於一區塊 之該等位元線可實施為垂直定向之二指間位元線群組。各 個別子線或位元線群組可由在陣列四側上的一個別解碼器/ 驅動器電路及一個別感應電路來服務。 如本文所使用,一列橫跨整個記憶體機架延伸(若不橫 跨整條)並包括許多字線。如本文所使用,” 一般跨越複數 個陣列區塊”之一匯流排或線包括幾乎跨越所有陣列區 塊,例如跨越全部但除最後區塊外(例如一給定匯流排不 耦合之一最後區塊)。此類匯流排或線可置放於陣列區塊 側,或可置放於此類a彳思體區塊上面或下面(即在一垂直 於一半導體基板之方向上)。 如本文所使用,&quot;將選定位元線耦合至一第一匯流排&quot;意 味著分別將各此類選疋位元線耦合至該第一匯流排之一對 應匯流排線。如本文所使用,字線(例如包括字線片段)與 位元線通常表示正父陣列線,且一般遵從此技術中的一普 通假設,即至少在一讀取操作週期驅動字線並感應位元 123178.doc -66 - 200826114 線。而且,如本文所使用,一&quot;全域線&quot;(例如一全域選擇 線)係跨越多個記憶體區塊之一陣列線,但不應得出任何 特定推論來暗示此類全域線必須橫跨—整個記憶體陣列或 實質上橫跨一整個積體電路。Fasoli application ("〇23-0056" application); US application No. 11/461,359 (attorney file number 023-0052), entitled "Incorporating data bus for memory array block selection" Memory array '', applied by Roy E. Scheuerlein, Luca G. Fasoli and Christopher J. Petti (''023-0052'' application); US application No. 11/46 1,372 (lawyer file number) 023-0057), Title 123178.doc -64- 200826114 π is used in the memory array block selection two data bus," by Roy Ε Scheuerlein, Luca G. Fasoli and Christopher J. Petti ( 〇23-0057" Application); US Application No. 11/46 1,362 (Attorney Docket No. 023-0053), titled π Hierarchical Bit Line Bias for Block Selectable Memory Arrays Busbars'', applied by Roy Ε. Scheuerlein and Luca G. Fasoli (, 〇 23-0053); and US Application No. 11/461, 376 (lawyer file number 023 - 0058), titled &quot;Using a hierarchical pattern for block selectable memory arrays The method of bit line bias bus bar &quot;, applied by Roy E. Scheuerlein and Luca G. Fasoli (&quot;023-0058&quot; application). It should be understood that the specific exemplary embodiments shown herein are always in a specific digit. The background of the example illustrates, for example, the number of decoded outputs, the number of decoder heads, the number of bus bars, the number of data busses, the number of array blocks in a memory frame, and the number of memory banks. Other teachings that conform to other design goals are implemented using the teachings of this disclosure. For clarity, not all of the conventional features of the &amp; scheme are described and illustrated. Most memory array designs have a relatively high Uniformity. For example, i usually includes the same number of memory cells per bit line. As another example, the number of primitive lines, word lines, array blocks, and even memory planes is often 2 integers. Power (eg '2n) to obtain simplification and efficiency of the decoding circuit. However, such regularity or consistency is of course not required for any particular embodiment of the invention. For example, For a different number of memory cells, the memory array may include three memory 123178.doc -65 - 200826114 body planes, and the word line segments in the first and last array blocks may be in the number of memory cells or bit line groups Any change in state and many other irregular changes to the general consistency of the memory array design, unless otherwise explicitly stated in the scope of the patent application, even as shown in the specific embodiments described herein, such general rules Sex should not be included in the meaning of any patent independent item. It should be understood that the indications of the top, left, bottom and right are merely illustrative terms for the four sides of a memory. The word line segments for a block may be implemented as horizontally oriented inter-finger word line segment groups, and the bit lines for a block may be implemented as vertically oriented two-finger bit lines Group. Each individual sub-line or group of bit lines can be served by a different decoder/driver circuit on the four sides of the array and a different sensing circuit. As used herein, a column extends across the entire memory frame (if not across the entire strip) and includes a number of word lines. As used herein, a bus or line that generally spans a plurality of array blocks includes almost all of the array blocks, such as spanning all but except the last block (eg, a given bus bar is not coupled to one of the last blocks) Piece). Such bus bars or wires may be placed on the side of the array block or may be placed above or below such a block (i.e., in a direction perpendicular to a semiconductor substrate). As used herein, &quot; coupling a selected location line to a first bus&quot; means coupling each such option bit line to one of the first bus bars corresponding to the bus bar, respectively. As used herein, a word line (eg, including a word line segment) and a bit line generally represent a positive parent array line, and generally conforms to a common assumption in the art that the word line is driven and sensed at least during a read operation cycle.元123178.doc -66 - 200826114 Line. Moreover, as used herein, a &quot;global line&quot; (e.g., a global selection line) spans one of a plurality of memory blocks, but no specific inference should be drawn to imply that such a global line must be horizontal Across the entire memory array or substantially across an entire integrated circuit.

如本文所使用,-讀取/寫入電路(例如一設定及讀取電 路)可用於一或多個資料位元,因此可耦合至一單一導 線,或可包括耦合至用於各分離資料位元之一資料匯流排 之各匯流排線的一分離此類讀取/寫入電路。 如本文所使用,一”資料匯流排”或資料匯流排&quot;片段,,至 少多次傳遞資料相依資訊,但不必始終如此。例如,此類 資料匯流排可針對特定操作模式來在此類資料匯流排之各 匯流排線上傳遞相同偏壓資訊。如本文所使用,一,,全域,, 匯流排可橫跨多個陣列區塊,但不必橫跨(或跨越)整個記 憶體陣歹。例%,此類全域匯流排可橫冑記憶體機架,而 不一定跨越一整個記憶體條。適當時,一,,資料電路”可包 括一讀取/寫入電路、一設定電路、一重置電路、一讀取 電路或一程式化電路之一或多個者或任一組合。 如本文所使用,’’選定”線(例如在一陣列區塊内的選定位 元線)對應於由一多頭解碼器電路同時選定並各耦人至一 對應匯流排線之此類位元線。此類位元線還可以或不可以 由資料或I/O電路來選定以實際執行一給定讀取、程式 化、設定、重置或抹除操作。例&gt;,若一 16頭行解碼r同 時”選擇&quot;並將16位元線耦合至一給定匯流排(例如SELN匯 流排)’則其涵蓋沒有任何位元線、一位元線、—個以上 123178.doc -67- 200826114 位元線或此16位元線群組之全部位元線可實際上接收一適 用於給定操作模式之選定偏壓條件,而剩餘位元線可接收 未&amp;疋偏壓條件。此類匯流排可說明為一&quot;資料相依&quot;匯 流排。在其他具體實施例中,可能存在一個以上此類&quot;選 定&quot;偏壓條件在一給定匯流排上傳遞,例如在二同時選定 記憶體單元將程式化成不同資料狀態時。 如本文所使用,—被動元件記憶體陣列包括複數個2端 子。己fe、體早7L,各連接於一相關聯乂線(例如字線)與—相 關聯Y線(例如位元線)之間。此類記憶體陣列可以係—二 、M平面)陣列或可以係一具有一個以上記憶體單元平面之 2維陣列。此類記憶體單元具有—非線性導電率,其中在 一反向方向(例如,從陰極至陽極)上的電流低於在—正向 方向上的電流。一被動元件記憶體陣列可以係一一次可程 式化(即一二欠寫入)記憶體陣列或一讀取/寫入(即多次寫入) 記,體陣列。此類被動元件記憶體陣列可—般視為具有一 a方向上引導電流之電流操縱元件與另一能夠改變其狀 心,組件(例如一溶絲、一反熔絲、一電容器、一電阻元 ^ ㈣元件時’該記憶體元件之程式化狀 態可藉由感應電流或電壓降來讀取。 在各圖中各種陣列線之方向性僅方便用於簡化陣列中二 交又線群組之說明。如本文所使用,一積體電路記憶體陣 列係—單石積體電路結構,而非一個以上封裝在一起或 密近接之積體電路。 、 文中方塊圖可使用一連接區塊之單一節點之術語來說 123178.doc -68- 200826114 明。雖然如此,作庫捧紐 1一應暸解,在背景要求時,此類”節點,,可 實際上表示用於傳遞一 #八 刀k戒之一對節點,或可表示用 於載送若干相關信號哎用於 、、, 一 A用於载迗形成一數位字或其他多位 兀仏5虎之複數個信號的客加八 ^ 观的夕個分離導線(例如匯流排)。 儘管一般假定電路及眚_ 貫體、、、°構,但應完全認識到,在現 代半導體設計及製造中,實體結構及電路可採用適用於後 續設計、測試或製造階段以及所產生製造半導體積體電路 之電腦可讀取描述性形式來具體化。因此,關於傳統電路 或結構之线符合其特定語言,可理解為其電腦可讀取編 馬及表不不响肷入於媒體内或是組合適當讀取器以允許 對應電路及/或結構之製造、測試或設計精細化。本發明 係還涵蓋以包括電路、包括此類電路之封裝模組、利用此 類電路及/或模組及/或其他記憶體裝置之系統、相關操作 方法、用於製造此類電路之相關方法、及此類電路及方法 之電腦可讀取媒體編碼’均如本文所述及如隨时請專利 範圍所定義。如本文所使用,—電腦可讀取媒體至少包括 磁碟、磁帶或其他磁性、光學、半導體(例如快閃記憶 卡、ROM)或電子媒體及一網路、有線、無線或其他通信 媒體。一電路之一編碼可包括電路圖資訊、實體佈局資 訊、行為模擬資訊及/或可包括可表示或傳達該電路之任 何其他編碼。 前述詳細說明僅已說明本發明之許多可行實施方案之— 些。由此原因,希望此詳細說明僅用以說明,而非用以限 制。可基於本文所提出之說明來進行本文所揭示之具體實 123178.doc -69- 200826114 施例之變化及修改而不脫離本發明之範疇及精神。僅希望 隨附申請專利範圍(包括全部等效内容)定義本發明之範 轉°而且,上述具體實施例明確涵蓋以單獨以及採用各種 組合來加以使用。因此,本文所述之其他具體實施例、變 化及改良不必脫離本發明之範_。 【圖式簡單說明】As used herein, a read/write circuit (eg, a set and read circuit) can be used for one or more data bits, and thus can be coupled to a single wire, or can include coupling to separate data bits. A separate read/write circuit of each of the bus bars of one of the data bus bars. As used herein, a “data bus” or data bus&quot;fragment, at least multiple data-receiving information, but not always. For example, such data buss can deliver the same bias information on each bus line of such data bus for a particular mode of operation. As used herein, a global domain can span multiple array blocks without having to span (or span) the entire memory matrix. For example, such a global bus can span a memory rack and does not necessarily span an entire memory bank. Where appropriate, the data circuit may comprise a read/write circuit, a set circuit, a reset circuit, a read circuit or a stylized circuit, one or more or any combination. The ''selected' lines (e.g., selected bit lines within an array block) are used to correspond to such bit lines that are simultaneously selected by a multi-head decoder circuit and each coupled to a corresponding bus bar. Such bit lines may or may not be selected by data or I/O circuitry to actually perform a given read, program, set, reset or erase operation. Example > If a 16-bit row decodes r "select" & couples a 16-bit line to a given bus (eg, SELN bus), then it covers no bit lines, one bit line, More than one of the 123178.doc -67-200826114 bit lines or all of the bit lines of the 16-bit line group may actually receive a selected bias condition suitable for a given mode of operation, while the remaining bit lines are receivable Not &amp; bias conditions. Such bus bars can be described as a &quot;data dependent&quot; bus. In other embodiments, there may be more than one such &quot;selected&quot; bias conditions given Passing on the bus, for example, when the two selected memory cells are programmed into different data states. As used herein, the passive component memory array includes a plurality of 2 terminals. The body is 61, and the body is 7L, each connected to a correlation. Between a connection line (such as a word line) and an associated Y line (such as a bit line). Such a memory array may be an array of two or M planes or may have a plane with more than one memory unit. Dimensional array. This type of memory Having a non-linear conductivity in which the current in a reverse direction (eg, from the cathode to the anode) is lower than the current in the -forward direction. A passive component memory array can be programmed once ( That is, one or two writes to the memory array or a read/write (ie, multiple writes), the body array. Such a passive component memory array can be generally regarded as having a current in the direction of the a direction. The operating element and another component capable of changing its shape, such as a solution wire, an anti-fuse, a capacitor, a resistor element (4), the stylized state of the memory element can be induced current or voltage drop To read. The directionality of the various array lines in each figure is only convenient for simplifying the description of the two-cross line group in the array. As used herein, an integrated circuit memory array is a single-span integrated circuit structure. Rather than one or more of the integrated circuits that are packaged together or closely connected. The block diagram in the text can be used in the terminology of a single node of a connected block. 123178.doc -68- 200826114. However, the library has a new one. Should When required by the background, such a "node" may actually be used to convey one of the eight knives or one of the nodes, or may be used to carry a number of related signals for use, , and an A for a separate conductor (such as a busbar) that forms a digital word or a plurality of other signals of a plurality of 虎5 tigers. Although it is generally assumed that the circuit and the 贯 、 、 、 、 It should be fully recognized, however, that in modern semiconductor design and fabrication, the physical structure and circuitry can be embodied in a computer readable descriptive form suitable for subsequent design, testing, or fabrication stages, as well as the resulting semiconductor integrated circuitry. Regarding the line of a conventional circuit or structure conforming to its specific language, it can be understood that its computer readable codec and table are not interrupted in the medium or combined with a suitable reader to allow the manufacture of the corresponding circuit and/or structure. Refine, test or design. The present invention also encompasses systems including circuits, package modules including such circuits, systems utilizing such circuits and/or modules and/or other memory devices, associated methods of operation, and methods for fabricating such circuits And computer readable media codes of such circuits and methods are as defined herein and as defined in the patent scope. As used herein, computer readable media includes at least a magnetic disk, magnetic tape, or other magnetic, optical, semiconductor (e.g., flash memory card, ROM) or electronic media and a network, wired, wireless, or other communication medium. One of the codes of a circuit may include circuit diagram information, physical layout information, behavioral simulation information, and/or may include any other code that may represent or communicate the circuit. The foregoing detailed description has set forth a few of the various embodiments of the invention. For this reason, it is intended that the detailed description be only illustrative and not restrictive. Variations and modifications of the specific embodiments disclosed herein may be made without departing from the scope and spirit of the invention. It is intended that the scope of the invention, including the claims Therefore, other specific embodiments, changes, and improvements described herein do not depart from the scope of the invention. [Simple description of the map]

參考附圖,習知此項技術者不僅可更理解本發明,還可 明白其許多目標、特徵及優點。 圖1係一記憶體陣列之一示意圖,說明選定及未選定字 線及位元線、及在一正向偏壓操作模式下的範例性偏壓條 件。 〃The present invention will be understood by those skilled in the art, and many of its objects, features and advantages are apparent. 1 is a schematic diagram of a memory array illustrating selected and unselected word lines and bit lines, and exemplary bias conditions in a forward bias mode of operation. 〃

圖2係圖1所示記憶體陣列之一示意圖, 偏壓操作模式下的範例性偏壓條件。 圖3係一字線解碼器電路之一示意圖, 壓操作條件下的範例性條件。 圖4係一字線解碼器電路之一示意圖, 壓操作條件下的範例性條件。 圖5係一位元線解碼器電路之一示意圖 偏壓操作條件下的範例性條件。 圖6係一位元線解碼器電路之一示意圖 偏壓操作條件下的範例性條件。 圖7係一字線解石民 尺鮮碼态電路之一示意圖, 他具體實施例在一及A Μ广 反向偏壓操作條件下的 圖8係一位元線解 鮮碼為電路之一示意圖 但ό兄明在一反向 包括在一正向偏 包括在一反向 偏 ,包括在一正向 ,包括在一反向 包括對於特定其 範例性條件。 ’包括對於特定 123178.doc -70- 200826114 其他具體實施例在一反向偏壓操作條件下的範例性條件。 圖9係一具有雙解碼源極選擇匯流排之字線解碼器電路 之示思圖’包括在一用於重置程式化之反向偏壓操作條 件下的範例性條件。 圖10係一具有資料相依源極選擇匯流排之位元線解碼器 電路之一示意圖,包括在一用於重置程式化之反向偏壓操 作條件下的範例性條件。 圖11係描述一包括一三維記憶體陣列之範例性積體電路 之一方塊圖,且該積體電路包括在該陣列一側的一全域列 解碼器與同時在該陣列頂部及底部的一對行解碼器。 圖12係表示依據本發明之特定具體實施例之一三維記憶Figure 2 is a schematic illustration of one of the memory arrays of Figure 1, with exemplary bias conditions in a bias mode of operation. Figure 3 is a schematic diagram of a word line decoder circuit, exemplary conditions under pressure operating conditions. Figure 4 is a schematic diagram of a word line decoder circuit, exemplary conditions under pressure operating conditions. Figure 5 is a schematic diagram of one of the one-line decoder circuits. Example conditions under bias operating conditions. Figure 6 is a diagram of one of the one-line decoder circuits. Example conditions under bias operating conditions. FIG. 7 is a schematic diagram of a fresh-coded circuit of a word line solution of the stone, and FIG. 8 is a schematic diagram of a single-line resolving code as a circuit under the conditions of A and A wide reverse bias operation. ό 明 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一 在一</ RTI> includes exemplary conditions under a reverse bias operating condition for other specific embodiments of 123178.doc-70-200826114. Figure 9 is a diagrammatic representation of a word line decoder circuit having dual decoded source select busses&apos; including exemplary conditions for resetting the stylized reverse bias operating conditions. Figure 10 is a schematic illustration of a bit line decoder circuit having data dependent source select busses, including exemplary conditions for resetting stylized reverse bias operating conditions. Figure 11 is a block diagram showing an exemplary integrated circuit including a three-dimensional memory array, and the integrated circuit includes a global column decoder on one side of the array and a pair at the top and bottom of the array simultaneously Line decoder. Figure 12 is a diagram showing a three-dimensional memory according to a specific embodiment of the present invention.

交錯的字線片段,其中至用於一 區塊之該等字線片段之 半的垂直連接係在該區塊左側,而至用於該區塊之該等字 線片段之另一半的垂直連接係在該區塊右側。此外,來自 二相鄰區塊之一字線片段共用各垂直連接。 圖13係一三維圖,其描述符合圖12所示者之特定具體實 施例之一三維記憶體陣列之一部分, 陣列區塊之各區塊内一個別字媿K jn ’並說明藉助至二相鄰Interleaved word line segments, wherein a vertical connection to a half of the word line segments for a block is to the left of the block, and a vertical connection to the other half of the word line segments for the block It is on the right side of the block. In addition, word line segments from one of the two adjacent blocks share the vertical connections. Figure 13 is a three-dimensional view depicting a portion of a three-dimensional memory array in accordance with a particular embodiment of the embodiment shown in Figure 12, a block 愧K jn ' in each block of the array block and illustrating the use of two-phase adjacent

線耦合至與該記憶體機架相關聯之 且各機架包複數個記憶體陣 同時選定’各將其個別位元 聯之二資料匯流排之一個別 123178.doc -71 - 200826114 者。 —:係、心德體機架之一方塊圖,說明另一配 二;體機架相關聯之二資料匯流排之一個別者。 二陣列區塊係顯示為同時選 酉己置,其令 M %fA ^ ^ . 將/、個別位元線耦合至 機架相關聯之二資料匯流排之-個別者。 -圖係-記憶體機架之一方塊圖,說明另一配中 一陣列區塊係顯示為同時 /、 M各將其個別位元線耦合至 ,、垓圯L體機架相關聯之二 m^μ^ ^ ^ 貝卞十匚机排之一個別者,該等 匯抓排係置放於該等記憶體陣列區塊之相同側上。 圖18係一記憶體機架之一方塊圖,說明另一配置,其中 一不相鄰陣列區塊待顯 盔 時選定’各將其個別位元線 l至與該記憶體機架相關聯之二資料匯流排之一個別 者0 圖19係-記憶體機架之一部分之一方塊圖,說明—範例 性階層式解竭配置用於在該等源極選擇匯流排上提供適當 條件用於選定及未選定陣列區塊。 田 _-記憶體機架之一部分之一方塊圖,f兒明另—範 例性階層讀碼g&amp;置料在該等源極選擇m卜上提供適 *條件用於選定及未選定陣列區塊。 圖21係一記憶體機架之一部分之一方塊圖,說明另一範 例性階層式解碼配置用於在該㈣極選擇匯流排上提供適 當條件用於選定及未選定陣列區塊。 ” 123178.doc -72- 200826114 圖22係一記憶體機架之一部分之一方塊圖,說明另一範 例性階層式解碼配置用於在該等源極選擇匯流排上提供適 當條件用於選定及未選定陣列區塊。 圖23係一資料電路之一方塊圖,其包括用於本文所述各 種具體實施例之一讀取感應放大器、一設定驅動器、及一 重置驅動器。The line is coupled to a memory array associated with the memory rack and each of the plurality of memory arrays is simultaneously selected 'one of the two data busses each of which has its individual bits connected to each other 123178.doc -71 - 200826114. —: A block diagram of the system and the body of the body, indicating the other two; one of the two data bus associated with the body frame. The two-array block system is shown as being simultaneously selected, which causes M %fA ^ ^ to couple /, individual bit lines to the individual of the data bus associated with the rack. - Figure-block diagram of a memory rack, indicating that another array of array blocks is displayed as simultaneous /, M each couples its individual bit lines to, and the L-body rack is associated with m^μ^ ^ ^ One of the individual rows of the Bellow Ten Thousand Machines, which are placed on the same side of the memory array block. Figure 18 is a block diagram of a memory rack illustrating another configuration in which a non-adjacent array block is selected to have its individual bit lines 1 associated with the memory rack when the helmet is to be helmeted. One of the two data bus rows 0 Figure 19 is a block diagram of one of the memory racks, illustrating that an exemplary hierarchical depletion configuration is used to provide appropriate conditions for selection on the source selection busbars And the array block is not selected. Field _-a block diagram of one of the memory racks, f-Make another - the exemplary stratum reading code g &amp; the stock provides the appropriate conditions on the source selection m b for selected and unselected array blocks . Figure 21 is a block diagram of one portion of a memory rack illustrating another exemplary hierarchical decoding configuration for providing suitable conditions for selected and unselected array blocks on the (four) pole select bus. 123178.doc -72- 200826114 Figure 22 is a block diagram of one portion of a memory rack illustrating another exemplary hierarchical decoding configuration for providing appropriate conditions for selection and selection on the source select busses. Figure 23 is a block diagram of a data circuit including a read sense amplifier, a set driver, and a reset driver for various embodiments described herein.

圖24係一範例性重置電路之一方塊圖,包括透過一選定 記憶體單元之重置路徑及該等字線及位元線選擇路徑之一 描述。 不同圖式中使用相同參考符號指示相似或相同的項目 【主要元件符號說明】 100 範例性被動元件記憶體陣列 101 被動元件記憶體單元 102 字線 103 被動元件記憶體單元 104 字線 105 被動元件記憶體單元 106 位元線 107 被動元件記憶體單元 108 位元線 152 列解碼器 153 電源節點 154 電源節點 155 輸出/節點 123178.doc -73- 200826114 156 反相器 157 多工器 158 解碼輸出 159 輸出 160 反相器 161 多工器 162 解碼輸出 164 節點 167 匯流排 168 匯流排線 171 PMOS電晶體 172 NMOS電晶體 173 PMOS電晶體 174 NMOS電晶體 175 PMOS電晶體 176 NMOS電晶體 177 PMOS電晶體 178 NMOS電晶體 181 未選定字線 183 未選定字線 200 範例性偏壓條件 202 行解碼器 203 電源節點 204 電源節點 123178.doc -74- 200826114 205 輸出 206 反相器 207 多工器 208 解碼輸出 209 輸出 210 反相器 211 多工器 212 解碼輸出 214 節點 217 匯流排 218 匯流排線 221 PMOS電晶體 222 PMOS電晶體 223 PMOS電晶體 224 NMOS電晶體 225 PMOS電晶體 226 NMOS電晶體 227 PMOS電晶體 228 NMOS電晶體 231 位元線 233 未選定位元線 300 範例性記憶體陣列 302 雙列解碼器 304 雙列解碼器 -75· 123178.doc 200826114 306 記憶陣列區塊 308 記憶陣列區塊 310 垂直連接 312 位元線電路區塊 314 位元線電路區塊 315 位元線電路區塊 316 位元線電路區塊 318 記憶體”條π 320 記憶體”條π 322 位元線 324 位元線 332 記憶體區塊 333 位元線 334 記憶體區塊 335 位元線 336 字線片段 337 字線片段 338 字線片段 339 垂直連接 340 垂直連接 342 字線片段 344 垂直連接 352 解碼輸出 353 解碼輸出 123178.doc -76- 200826114 358 359 360 361 362 363 370 371 372 373 374 375 376 377 378 379 380 381 382 400 402 404 406 407 垂直連接 垂直連接 字線片段 字線片段 字線片段 字線片段 記憶體陣列 第一條 第二條 頂部資料匯流排 記憶體陣列區塊 記憶體陣列區塊 選定字線 列 底部資料匯流排 底部行解碼器電路 頂部行解碼電路 頂部位元線選擇區塊 底部位元線選擇區塊 記憶體機架 第一資料匯流排 第二資料匯流排 奇數陣列區塊 偶數陣列區塊 123178.doc -77- 200826114 408 410 412 420 422 424 426 427 430 432 440 442 444 446 447 448 449 450 454 460 462 464 466 468 位元線選擇區塊 粗體箭頭 粗體箭頭 記憶體機架 資料匯流排 第二資料匯流排 第一陣列區塊 第二陣列區塊 粗體箭頭 粗體箭頭 記憶體機架 第一資料匯流排 第二資料匯流排 第一陣列區塊 陣列區塊 第二位元線選擇區塊 第一位元線選擇區塊 粗體箭頭 粗體箭頭 記憶體機架 陣列區塊 陣列區塊 上部資料匯流排 下部資料匯流排 123178.doc -78- 200826114Figure 24 is a block diagram of an exemplary reset circuit including a reset path through a selected memory cell and one of the word line and bit line select paths. The same reference numerals are used in the different drawings to indicate similar or identical items. [Main element symbol description] 100 Exemplary passive element memory array 101 Passive element memory unit 102 Word line 103 Passive element memory unit 104 Word line 105 Passive element memory Body unit 106 bit line 107 passive element memory unit 108 bit line 152 column decoder 153 power node 154 power node 155 output / node 123178.doc -73- 200826114 156 inverter 157 multiplexer 158 decode output 159 output 160 Inverter 161 multiplexer 162 Decode output 164 Node 167 Bus 168 Bus bar 171 PMOS transistor 172 NMOS transistor 173 PMOS transistor 174 NMOS transistor 175 PMOS transistor 176 NMOS transistor 177 PMOS transistor 178 NMOS Transistor 181 Unselected Word Line 183 Unselected Word Line 200 Exemplary Bias Condition 202 Row Decoder 203 Power Node 204 Power Node 123178.doc -74- 200826114 205 Output 206 Inverter 207 Multiplexer 208 Decode Output 209 Output 210 inverter 211 multiplexer 212 decode output 214 Node 217 Bus 218 Bus 221 PMOS transistor 222 PMOS transistor 223 PMOS transistor 224 NMOS transistor 225 PMOS transistor 226 NMOS transistor 227 PMOS transistor 228 NMOS transistor 231 bit line 233 unselected locator Line 300 Exemplary Memory Array 302 Dual Column Decoder 304 Dual Column Decoder - 75 · 123178.doc 200826114 306 Memory Array Block 308 Memory Array Block 310 Vertical Connection 312 Bit Line Circuit Block 314 Bit Line Circuit Area Block 315 bit line circuit block 316 bit line circuit block 318 memory "strip π 320 memory" strip π 322 bit line 324 bit line 332 memory block 333 bit line 334 memory block 335 Bit Line 336 Word Line Segment 337 Word Line Segment 338 Word Line Segment 339 Vertical Connection 340 Vertical Connection 342 Word Line Segment 344 Vertical Connection 352 Decode Output 353 Decode Output 123178.doc -76- 200826114 358 359 360 361 362 363 370 371 372 373 374 375 376 377 378 379 380 381 382 400 402 404 406 407 Vertical connection vertical connection word line segment word line segment word line Segment word line segment memory array first second second top data bus memory array block memory array block selected word line column bottom data bus bottom line decoder circuit top row decoding circuit top bit line selection area Block bottom bit line selection block memory frame first data bus second data bus array odd array block even array block 123178.doc -77- 200826114 408 410 412 420 422 424 426 427 430 432 440 442 444 446 447 448 449 450 454 460 462 464 466 468 Bit Line Selection Block Bold Arrow Bold Arrow Memory Rack Data Bus Second Data Bus First Array Block Second Array Block Bold Arrow Bold Arrow memory rack first data bus second data bus first array block array block second bit line selection block first bit line selection block bold arrow bold arrow memory rack array The upper data bus of the upper block of the block array block is the data bus 123178.doc -78- 200826114

500 匯流排配置 502 記憶體陣列區塊 504 記憶體陣列區塊 506 記憶體陣列區塊 508 耦合電路 532 耦合電路 533 耦合電路 535 耦合電路 536 匯流排 550 匯流排配置 552 耦合電路 554 耦合電路 601 I/O邏輯 602 匯流排 603 匯流排 604 寫入鎖存器區塊 605 讀取鎖存器 606 匯流排 607 匯流排 608 控制邏輯 609 匯流排 610 匯流排 611 匯流排 612 控制線 123178.doc -79- 200826114 613 614 615 616 617 632 633 634 、 635 636 637 638 639 讀取感應放大器 設定驅動器 重置驅動器 SELB匯流排 SELN匯流排 信號 位元線預充電(BLP)限流電路 重置限制電路 SELN匯流排線 位元線選擇路徑 控制信號 選定記憶體單元 字線選擇路徑 123178.doc -80-500 bus arrangement 502 memory array block 504 memory array block 506 memory array block 508 coupling circuit 532 coupling circuit 533 coupling circuit 535 coupling circuit 536 bus bar 550 bus bar configuration 552 coupling circuit 554 coupling circuit 601 I / O logic 602 bus 603 bus 604 write latch block 605 read latch 606 bus 607 bus 608 control logic 609 bus 610 bus 611 bus 612 control line 123178.doc -79- 200826114 613 614 615 616 617 632 633 634 , 635 636 637 638 639 Read sense amplifier set driver reset driver SELB bus SELN bus signal bit line precharge (BLP) current limit circuit reset limit circuit SELN bus line position The line selection path control signal selects the memory unit word line selection path 123178.doc -80-

Claims (1)

200826114 十、申請專利範圍: L 一種積體電路,其包含·· -二己憶體陣列,其具有複數個陣列區塊,各陣列區塊 包含子線及位元線,· -弟-全域匯流排,其一般跨越該複數個陣列區塊, 用於有時將一選定區堍之登 • 尾之選疋位兀線耦合至個別資料電 路;及 每陣列區塊的一個別第一匯流排片段,其用於在一第 -#作模式期間將—適合於該第一操作模式之第一未選 =元線偏屢條件傳遞至-選定區塊之未選定位元線: 並將一適合於該第一操作 2. 叭八炙弟一未選疋位元線偏壓 條件麵5至未選定陣列區塊之未選定位元線。 如請求項1之積體電路,其中·· 4第-操作模式包含_設定操作模式; 二於該第一操作模式之該第一未選定位元線偏壓條 件包含一未選定位元線電壓;以及 ^於該第-操作模式之該第二未選w偏壓條 件包含一浮動條件。 3·如請求項2之積體電路,其中: 在弟一才呆作核式下,一選定陣列區塊之該個別第一 匯流排片段_合以將適合於該第二操作模式在個別第 =排片段匯流排線上之個別資料相依偏㈣件傳遞 至該選疋區塊之選定位元線。 4·如請求項3之積體電路,其中: 123178.doc 200826114 該第一全域匯流排包含耦 ^ 取/寫入匯流排。 至一項取/寫入電路之一讀 5. 如請求項4之積體電路,苴一 /、心一步包含: 一耦合至一第二資料電路 其中各區塊包括,c流排; 別第-匯流排片段輕合至該第於有時將該個 浮動。 王域匯流排,有時使之 τ 6·如請求項5之積 在該第一操作模式下,該 線係在適合”第…二一王域匯流排之各匯流排 條件下二式之第一未選定位元線偏壓 條件下偏壓;以及 繞作模式下’該第一全域匯流排之各匯流排 :、適合於該第二操作模式之未選定位元線偏壓條 件下偏壓。 7·如請求項3之積體電路,其進一步包含·· 一全域偏壓線,其耦合至一偏壓電路; 八各區塊包括一個別耦合電路,其用於有時將該個 別弟一匯流排片段之各匯流排線耦合至該全域偏壓線, 在此期間該全域偏壓線傳遞一適合於該第一操作模式之 未選定位元線偏壓電壓。 8·如請求項7之積體電路,其進一步包含: 複數個輕合電路,其用於有時將用於各個別陣列區塊 之該等個別第一匯流排片段耦合在一起,以形成跨越該 複數個陣列區塊之一單一邏輯匯流排,該等耦合一起的 123178.doc 200826114 第一匯流排片段係# 獅口至邊弟二資料電路。 9.如請求項8之積體電路,其中: 用於各個別陳万π W鬼之該等個別第一匯流排片段係在 ^弟一刼作模式期間耦合在一起。 10·如請求項8之積體電路,其中·· 在邊第二操作模式, t亥弟一全域匯流排之各匯流排 線係在一適合於琴筮_ 一刼作模式之未選定位元線電壓條 件下偏壓;以及 :該等第一及第二操作模式下,該全域偏壓線傳 遞:適合於該第一操作模式之未選定位元偏壓電壓。 11·如凊求項7之積體電路,其中: 各區塊包括-個別麵合電路,其用於有時將該個別第 -匯流排片段耦合至該第二全域匯流排。 I2·如請求項11之積體電路,其中: 在該第二操作模式下,嗲坌_ μ第一王域匯流排之各匯流排 綠係在一適合於該第二择作媪 之未敎μ線電壓條 件下偏壓;以及 在該等第一及第二操作模式_ 、八一者下,该全域偏壓線傳 遞-適合於該第-操作模式之未選定位元偏壓電壓。 13·如請求項7之積體電路,其中: 該第一全域匯流排包含麵合至該等個別資料電路之一 全域選擇匯流排;以及 該記憶體陣列進-步包含每陣列區塊—個別第二匿流 排片段,其用於在該第一操作模式期間,將—適合於: 123178.doc 200826114 =操作模式之資料相依位元線偏㈣件 區塊之撰它&gt; - μ 、王 選疋 -適合於ρ !用於在該第二操作模式期間,將 至未^ 作模式之未選定位元線㈣條件傳遞 至未、疋陣列區塊之未選定位元線。 14.如請求項13之積體電路,其中·· 各區塊包括一個別耦合電路,其用於有時將個別第二 匯流排片段輕合至該全域選擇匯流排,並在特定其他時 侯耦5至忒全域偏壓線,並在其他時候使之浮動; 在該第一操作模式下,該全域偏壓線傳遞一適合於該 第一操作模式之未選定位元偏壓電壓;以及 在該第二操作模式下,該全域偏壓線傳遞一適合於該 第二操作模式之未選定位元偏壓電壓。 15.如請求項1之積體電路,其中: 該記憶體陣列包含一三維記憶體陣列,其具有在二位 元線層上的位元線;以及 在一個別陣列區塊之二位元線層上的位元線係同時耦 合至個別匯流排片段。 16 ·如請求項15之積體電路,其中·· 各記憶體單元包含一可逆電阻器元件。 17. 如請求項16之積體電路,其中: 該可逆電阻器元件包含一過渡金屬氧化物。 18. 如請求項16之積體電路,其中: 各a己憶體早凡包含與一二極體串列的一可逆電阻is凡 件0 123178.doc 200826114 19 · 一種編碼如請求項1之積體電路之電腦可讀取媒體。 20. —種封裝模組,其包括請求項1之積體電路。 21· —種積體電路,其包含·· 一記憶體陣列,其具有複數個陣列區塊,各陣列區塊 包含子線及位元線;200826114 X. Patent application scope: L An integrated circuit, which comprises a double-resonant array, which has a plurality of array blocks, each array block includes a sub-line and a bit line, and the ---the global convergence a row, which generally spans the plurality of array blocks, for coupling a selected one of the selected regions to an individual data circuit; and a first bus segment of each array block And for transmitting the first unselected = element line partial condition suitable for the first mode of operation to the unselected location line of the selected block during a first-# mode: and adapting one The first operation 2. A bit of the unselected positioning element line 5 to the unselected positioning element line of the unselected array block. The integrated circuit of claim 1, wherein the fourth operation mode includes a set operation mode; and the first unselected positioning element line bias condition of the first operation mode includes an unselected positioning element line voltage And the second unselected w bias condition of the first mode of operation includes a floating condition. 3. The integrated circuit of claim 2, wherein: in the case of the younger one, the individual first bus segment of the selected array block is combined to be suitable for the second mode of operation in the individual = The individual data on the bus segment line is transmitted to the selected location line of the selected block. 4. The integrated circuit of claim 3, wherein: 123178.doc 200826114 The first global bus includes a coupling/writing bus. Read to one of the fetch/write circuits. 5. The integrated circuit of claim 4, the first step, includes: a coupling to a second data circuit, wherein each block includes a c-flow row; - The bus segment is lightly coupled to the first and sometimes floats. The king domain bus, sometimes making it τ 6 · If the product of the request item 5 is in the first mode of operation, the line is in the second type of the busbar condition suitable for the "..." two king domain busbars a bias voltage under unselected positioning element line bias conditions; and a bus bar of the first global bus bar in the winding mode: biasing under unselected positioning element line bias conditions suitable for the second mode of operation 7. The integrated circuit of claim 3, further comprising: a global bias line coupled to a bias circuit; each of the eight blocks includes a separate coupling circuit for sometimes the individual Each bus bar of the bus-strip segment is coupled to the global bias line during which the global bias line delivers an unselected locating line bias voltage suitable for the first mode of operation. The integrated circuit of 7 further comprising: a plurality of light combining circuits for coupling the individual first bus bar segments for the respective array blocks to form across the plurality of array regions One of the blocks of a single logic bus, the coupling one 123178.doc 200826114 The first bus segment is #狮口至边弟二数据电路. 9. The integrated circuit of claim 8, wherein: the individual first confluence for each Chen Wan π W ghost The segment segments are coupled together during the mode of the brother-in-law. 10. The integrated circuit of claim 8, wherein in the second mode of operation, each bus line of the global bus is tied to a bias voltage suitable for the unselected positioning element line voltage condition of the hammer mode; and: in the first and second modes of operation, the global bias line transmission: suitable for the first mode of operation The locating element bias voltage is not selected. 11. The integrated circuit of claim 7, wherein: each block includes an individual facet circuit for sometimes coupling the individual first bus bar segment to the second The global bus is as follows: I2. The integrated circuit of claim 11, wherein: in the second mode of operation, each of the 汇_μ first king-domain busbars is adapted to the second option Unbiased under the condition of the u line voltage; and in the And the second operation mode _, the eight-one, the global bias line transmission - the unselected locator bias voltage suitable for the first operation mode. 13. The integrated circuit of claim 7, wherein: A global bus includes a global selection bus that is coupled to one of the individual data circuits; and the memory array further includes each of the array blocks - an individual second bus segment for use in the first operation During the mode, it will be - suitable for: 123178.doc 200826114 = The data of the operating mode depends on the bit line offset (four) piece of the block it is written - μ, Wang Xuan - suitable for ρ ! used in the second mode of operation During this period, the unselected positioning element (4) condition to the un-mode is passed to the unselected positioning element of the un-arranged block. 14. The integrated circuit of claim 13, wherein each block comprises a separate coupling circuit for sometimes splicing individual second bus segments to the global selection bus and at a particular time Coupling 5 to the global bias line and floating it at other times; in the first mode of operation, the global bias line transmits an unselected locator bias voltage suitable for the first mode of operation; In the second mode of operation, the global bias line delivers an unselected locator bias voltage suitable for the second mode of operation. 15. The integrated circuit of claim 1, wherein: the memory array comprises a three-dimensional memory array having bit lines on a two bit line layer; and a two bit line in a different array block The bit line on the layer is simultaneously coupled to the individual bus segment. 16. The integrated circuit of claim 15, wherein each memory cell comprises a reversible resistor element. 17. The integrated circuit of claim 16, wherein: the reversible resistor element comprises a transition metal oxide. 18. The integrated circuit of claim 16, wherein: each a memory element comprises a reversible resistance of a string of diodes. 123 123.doc 200826114 19 · A product coded as claim 1 The computer of the body circuit can read the media. 20. A package module comprising the integrated circuit of claim 1. An integrated circuit comprising: a memory array having a plurality of array blocks, each array block comprising a sub-line and a bit line; 耦合構件,其用於在一第一操作模式下,藉由一般跨 越該複數個陣列區塊之—第—全域匯流排將—選定區塊 之選定位元線耦合至個別資料電路; 一麵合構件,其用於將二選定及未選定陣列區塊之未選 定位元線耦合至與各個別陣列區塊相關聯的一個別第一 匯流排片段; 塊相關聯之個別 操作模式之第一 傳遞構件,其用於在與該選定陣列區 第一匯流排片段上傳遞一適合於該第一 未選定位元線偏壓條件;以及 傳遞構件,其用於在與未選定 夕U塊相關聯之個別 二上傳遞-適合於該第-操作模式之第二 未選定位元線偏壓條件。 22.如請求項21之積體電路,其中: 該第一操作模式包含—設定操作模式; 適合於該第一操作模式之該 件包含-未選定位元線電壓;以及未…元線偏壓條 選定位元線偏壓條 適合於該第一操作模式之該第二 件包含一浮動條件。 23·如請求項21之積體電路, 其進一步包含在一 第二操作模 123178.doc 200826114 式下: 耦合構件,其用於將一選定陣列區塊之—或多個選定 位元線耦合至與該選定陣列區塊相關聯之個別第一匯流 排片段; ^ 傳遞構件’其用於在與該選定陣列區塊相關聯之個別 第-匯流排片段上傳遞一適合於該第二操作模式之個別 資料相依偏壓條件;以及 耦合構件,其用於將該選定陣列區塊之未選定位元線 耦合至该第一全域匯流排。 24. 25. 26. 27. 如請求項23之積體電路,其進一步包含: 在該第二操作模式下,用於將與該選定陣列區塊相關 聯之該個別第一匯流排片段耦合至一資料電路之構件。 如請求項23之積體電路,其中: 該第二操作模式包含一重置操作模式; 耦合至該選定陣列區塊之選定位元線的適合於該第二 操作模式之該等個別資料相依偏壓條件包含一使一選定 位元線重置的第一值與一使一選定位元線不改變的第二 值。 如請求項23之積體電路,其中: 該第一全域匯流排包含耦合至一讀取/寫入電路之一讀 取/寫入匯流排。 W 如請求項26之積體電路,其進一步包含·· 耦5構件其用於有時將用於一陣列區塊之個別第一 Ε μ排片奴耦合至一第二全域匯流排,該第二全域匯流 123178.doc 200826114 排自身係耦合至一第二杳极# &amp; 一貝科電路’並有時用於使該個別 第一匯流排片段浮動。 28. 如請求項23之積體電路,其進_步包含: 耦合構件,其用於有時將該個別第一匯流排片段之各 匯流排片段耗合至一全域偏壓線;以及 傳遞構件’其用於有時在該全域偏壓線上傳遞一適合 於該第-操作模式之未選定位元線偏壓電壓。 29. 如請求項28之積體電路,其進一步包含: 躺合構件,其用於有時 等將用於各個別陣列區塊之該等 個別第一匯流排片段耦人 個陣列區… 起,以形成一跨越該複數 個陣紅塊之早一邏輯匯流排;以及 耦合構件,其用於 耦合至$ m' 耦5 一起的第一匯流排片段 祸口至邊弟二資料匯流排。 30·如請求項28之積體電路,其進-步包含: 至二構用於有時將該個別第-匯流排片段麵合 王4罘_全域匯流排。 31·如請求項28之積體電路,其令: 弟王域匯流排包含;人5 ·&gt;&gt;· ·4* 全域選摆匕3耦&amp;至该等個別資料電路之— A k擇匯流排;以及 該積體電路進—步包 區塊的一個別贫_ 具用於精由母陣列 間,將適合於:第—匯广排片段’在該第-操作模式期 屋條件傳遞至=作模式之個別資料相依位元線偏 、王—選定區塊之遴一 二操作模^間,將線’並用於在該第 、、σ於该第二操作模式之未選定 123178.doc 200826114 位元線偏壓條件傳遞至未選定陣列區塊之未選定位元 線。 32. —種用於配合一記憶體陣列使用之方法,該記憶體陣列 具有複數個陣列區塊,各陣列區塊包括字線與位元線, 在一第一操作模式下,該方法包含: 藉由 般跨越該複數個陣列區塊之第一區域匯流 ί 排,將一選定區塊之選定位元線耦合至個別資料電路; 將選疋及未選定陣列區塊二者之未選定位元線耦合至 與各個別陣列區塊相關聯的一個別第一匯流排片段; 在與該選定陣列區塊相關聯之個別第一匯流排片段上 傳遞一適合於該第一操作模式筮 ._ ^ 7 ^ Α 你F揭八怎第一未選定位元線偏壓 條件;以及 在與未選定陣列區塊相關聯之個別第一匯流排片段上 傳遞-適合於該第一操作模式之第二未選定位元線偏壓 條件。 3 3 ·如請求項3 2之方法,其中: 適合於該第一操作模式之該第- 人… 弟一未選疋位元線偏壓條 件包含一淨動條件。 3 4 ·如請求項3 3之方法,其中·· 该第-操作模式包含一設定操作模式;以及 適合於該第一操作模式之該第一 件包含一去、登禾k疋位元線偏壓條 1于匕3 未選定位元線電壓。 35.如請求項32之方法,其進 下: 匕3在—弟二操作模式 123178.doc 200826114 將一選定陣列區塊之一 ^ ^ /夕個選疋位元線麵合至與該 k疋陣列區塊相關聯之個 ’、 ⑴第一匯流排片段; 在與該選定陣列區堍相 ^, 塊相關聯之個別第一匯流排片段上 傳遞一適合於該第二操 ,4, . ^ ^ 枳式之個別資料相依偏壓條 1干,以及 將該選定陣列區塊之去 匯流排。 ^疋立兀線耦合至該第一全域 36. 如請求項35之方法,其進_步包括: 在该第二操作模式下, 一 將,、5亥選疋陣列區塊相關聯之 個別第一匯流排片段勉人 又耦$至一資料電路。 37. 如請求項36之方法,其進一步包括: 在β亥第二操作模式下〜 將/、5亥選疋陣列區塊相關聯之 個別第一匯流排片段差 ^ 〇至一跨越該複數個陣列區塊之 匯流排。 3 8.如請求項35之方法,其中·· 該第二操作模式包含-重置操作模式; 至該選^陣列區塊之選m線的適合於該第二 插作拉式之該等個別資料相依偏壓條件包含—使一選定 線重置之第—值與一使一選定位元線不改變之第二 值。 39.如請求項35之方法,其中: 該第—全域匯流排包含搞合至—讀取/寫入電路之-讀 取/寫入匯流排。 後如請求項39之方法,其進一步包括: 123178.doc 200826114 =於-陣列區塊之個別第—匯流排片段輕合至 第二:二:匯流排’該第二全域匯流排自身係耦合至一 動路’並有時用於使該個別第-匯流排片段浮 動0 41. 如請求項4〇之方法,其進一步包含: 在該第一操作模式 M ^ y. 、以弟一王域匯流排之各匯流 排線在適合於該第一操 條件下偏壓;以1 式之弟―未選定位元線偏壓 排!Γ第二操作模式下,將該第一全域匯流排之各匯流 m合於㈣:操作料之未敎位域偏 件下偏壓。 42. 如請求項35之方法,其進一步包含: 有時將該個別第-匯流排片段之各匯流排線麵合至一 全域偏壓線;以及 在該等時候在該全域偏壓線上傳遞一適合於該第―操 作模式之未選定位元線偏壓電壓。 43. 如請求項42之方法,其進一步包含: 有時將用於各個別陣列區塊之該等個別第一匿流排片 段耦合在一起,以形成一跨越該複數個陣列區塊之單一 邏輯匯流排;以及 將該等輕合-起的第一匯流排片段麵合至該第二資料 匯流排。 ' 44·如睛求項43之方法,其進一步包含: 在該第二操作模式期間將用於各個別陣列區塊之該等 123178.doc -10 - 200826114 個別第一匯流排片段耦合在一起。 45.如請求項43之方法’其進_步包含: 在该第二操作模式下, 排線在-適合於該第二操作全域匯流排之各匯流 件下偏壓;以及 、$之未選定位70線偏壓條 在該等第一及第二操作模 ,^ ^ Α 、考下,在該全域偏壓後 上傳遞-適合於該第一操 錢深 壓。 乍挺式之未選定位元偏壓電 46·如請求項42之方法,其進_步包含: 有時將該個別第一匯流排 排。 片奴耦合至該第二全域匯流 47·如請求項46之方法,其進一步包含: 在該第二操作模式下,將 竹雀苐一全域匯流排之各匯流 排線在一適合於該第二操作楹 从 作枳式之未選定位元線偏壓條 件下偏壓;以及 在該等第一及第二操作模式— 飞一者下,在該全域偏壓線 上傳遞一適合於該第一掉竹播斗、 丄 呆1乍模式之未選定位元偏壓電 壓。 48.如請求項42之方法,其中: 該第一全域匯流排包含鉍人$ — μ , 3耦合至该等個別資料電路之一 全域選擇匯流排,以及 該方法進一步包含藉由每陣列F撿 可丨平外&amp;塊的一個別第二匯流 排片段’在該第-操作模式期間’將適合於該第一操作 模式之個別資料相依位元線偏壓條件傳遞至—選定區塊 123178.doc -11 - 200826114 之&amp;疋位元線,並用於在該第二操作模式期間,將一適 口於β第—操作模式之未選定位元線偏壓條件傳遞至未 選定陣列區塊之未選定位元線。 49·如請求項48之方法,其中: 在忒第一刼作模式期間傳遞至該選定陣列區塊之選定 位元線的該等個別資料相依偏壓條件包含-使-選定位 元線在其上操作之第-值與-使-選定位元線不改變之 第二值。a coupling member for coupling a selected positioning element line of the selected block to the individual data circuit by a first-global bus bar that generally spans the plurality of array blocks in a first mode of operation; a means for coupling unselected positioning element lines of the two selected and unselected array blocks to a different first bus line segment associated with each of the individual array blocks; a first pass of the individual operational modes associated with the block a means for transmitting a first unselected positioning element line bias condition on the first busbar segment with the selected array area; and a transfer means for associating with the unselected U-block Passing on the second two - a second unselected positioning element line bias condition suitable for the first mode of operation. 22. The integrated circuit of claim 21, wherein: the first mode of operation comprises - setting an operating mode; the component adapted to the first mode of operation comprises - unselected bit line voltage; and un... Selecting the positioning element line bias strip is suitable for the second piece of the first mode of operation to include a floating condition. 23. The integrated circuit of claim 21, further comprising a second operating mode 123178.doc 200826114: a coupling member for coupling a selected array block or a plurality of selected positioning elements to An individual first bus segment associated with the selected array block; ^ a transfer component 'for transmitting a suitable second mode of operation on the individual first bus segment associated with the selected array block The individual data is dependent on the bias condition; and a coupling member is operative to couple the unselected positioning element of the selected array block to the first global bus. 24. 25. 26. 27. The integrated circuit of claim 23, further comprising: in the second mode of operation, coupling the individual first busbar segment associated with the selected array block to A component of a data circuit. The integrated circuit of claim 23, wherein: the second mode of operation comprises a reset mode of operation; and the individual data coupled to the selected location line of the selected array block is adapted to the second mode of operation The pressing condition includes a first value for resetting the selected positioning element line and a second value for causing the selected positioning element line to not change. The integrated circuit of claim 23, wherein: the first global bus includes a read/write bus coupled to a read/write circuit. W. The integrated circuit of claim 26, further comprising: a coupling member for coupling an individual first Ε μ row of slaves for an array block to a second global bus, the second The global convergence 123178.doc 200826114 is coupled to a second drain # &amp; a Becco circuit' and is sometimes used to float the individual first bus segment. 28. The integrated circuit of claim 23, further comprising: a coupling member for sometimes consuming each bus bar segment of the individual first bus bar segment to a global bias line; and a transfer member 'It is used to sometimes pass an unselected positioning element line bias voltage suitable for the first mode of operation on the global bias line. 29. The integrated circuit of claim 28, further comprising: a lie member for sometimes equipping the individual first busbar segments for the respective array blocks into an array region... To form an early logic bus that spans the plurality of red patches; and a coupling member for coupling to the first bus segment of the $m' coupling 5 to the two-party data bus. 30. The integrated circuit of claim 28, wherein the further step comprises: to two configurations for sometimes merging the individual first bus bar segments with a global bus. 31. The integrated circuit of claim 28, wherein: the brother domain bus includes; the person 5 ·&gt;&gt;·4* the global selection of the 3 coupling &amp; to the individual data circuits - A k Selecting the bus bar; and the other circuit of the integrated circuit into the block block is used for the fine between the parent array, which will be suitable for: the first - the wide-range segment in the first operating mode至 ====================================================================================================== The 200826114 bit line bias condition is passed to the unselected location line of the unselected array block. 32. A method for cooperating with a memory array, the memory array having a plurality of array blocks, each array block comprising a word line and a bit line, and in a first mode of operation, the method comprises: Selecting a selected location element line of the selected block to the individual data circuit by traversing the first area of the plurality of array blocks, and selecting unselected positioning elements for both the selected and unselected array blocks a line coupled to a different first bus segment associated with each of the array blocks; transmitting a suitable one for the first mode of operation on the respective first bus segment associated with the selected array block _._^ 7 ^ Α You F unveil the first unselected locating element line bias condition; and pass on the individual first bus snippet segments associated with the unselected array block - the second unsuitable for the first mode of operation Select the positioning element line bias condition. 3: The method of claim 3, wherein: the first person suitable for the first mode of operation, the first unselected bit line bias condition comprises a net moving condition. 3. The method of claim 3, wherein: the first mode of operation comprises a set mode of operation; and the first component adapted to the first mode of operation comprises a de-emphasis The bead 1 is not selected to locate the line voltage at 匕3. 35. The method of claim 32, wherein: 匕3 in the second mode of operation 123178.doc 200826114, one of the selected array blocks is ^^ / 夕 疋 疋 线 线 与 与Array block associated with one, (1) first bus segment segment; in the selected first array region, the individual first bus segment segment associated with the block is passed a suitable for the second operation, 4, . ^ The individual data of the 相 type is dependent on the bias strip 1 and the de-sink of the selected array block. The 兀 兀 line is coupled to the first global domain 36. The method of claim 35, wherein the step 301 comprises: in the second mode of operation, an individual, associated with the 5th 疋 array block A bus segment is coupled to a data circuit. 37. The method of claim 36, further comprising: in the second mode of operation of the second mode, the individual first bus slice segments associated with the /, 5, and 疋 array blocks are crossed to the plurality of arrays The bus of the block. 3. The method of claim 35, wherein: the second mode of operation comprises a reset mode of operation; wherein the plurality of selected m lines of the selected array block are suitable for the second plug-in type The data dependent bias condition includes a first value that resets a selected line and a second value that does not change the selected positioning line. 39. The method of claim 35, wherein: the first global buffer includes a read/write bus that is coupled to the read/write circuit. The method of claim 39, further comprising: 123178.doc 200826114=The individual-busbar segments of the array block are lighted to the second: two: busbars' the second global busbar is itself coupled to a path 'and sometimes used to float the individual first bus segment 0. 41. The method of claim 4, further comprising: in the first mode of operation M ^ y. Each of the bus bars is biased under the first operating condition; the mode is biased by the unselected positioning element line of the type 1! In the second mode of operation, each of the first global busbars is merged with (4): the bias of the operating region is biased. 42. The method of claim 35, further comprising: sometimes combining the respective busbar lines of the individual first busbar segments to a global bias line; and transmitting one of the global bias lines at the time An unselected positioning element line bias voltage suitable for the first mode of operation. 43. The method of claim 42, further comprising: coupling the individual first bust segments for respective array blocks to form a single logic across the plurality of array blocks a bus bar; and the first bus segment that is lightly coupled to the second data bus. 44. The method of claim 43, further comprising: coupling the respective first bus segments for the respective array blocks during the second mode of operation. 45. The method of claim 43, wherein the second step comprises: in the second mode of operation, the cable is biased at each of the busbars suitable for the second operational global bus; and, $ is unselected The 70-line bias strip is transferred over the global bias after the first and second operating modes, ^^ 、, and is suitable for the first operating deep pressure. The unselected locating element bias voltage 46. The method of claim 42, wherein the step _ comprises: arranging the individual first bus at times. The method of claim 46, further comprising: in the second mode of operation, each of the bus bars of the global bus bar is adapted to the second Operating the 偏压 bias from the unselected locating element line bias condition; and in the first and second modes of operation - flying one, transmitting a suitable one on the global bias line Bamboo carrier, 丄 乍 1乍 mode unselected positioning element bias voltage. 48. The method of claim 42, wherein: the first global bus bar comprises a monk $ — μ , 3 coupled to one of the individual data circuits, the global selection bus, and the method further comprising An additional second bus segment 'in the first-operation mode' can pass the individual data-dependent bit line bias conditions suitable for the first mode of operation to the selected block 123178. Doc -11 - 200826114 &amp; bit line, and used to pass an unselected locating element line bias condition in the beta mode of operation to the unselected array block during the second mode of operation Select the positioning element line. The method of claim 48, wherein: the individual data dependent bias conditions passed to the selected positioning element of the selected array block during the first mode of operation include - enabling - selecting a location line in The first value of the upper operation and the - make-select positioning element line do not change the second value. 5〇·如請求項48之方法,其進一步包含: 有時將用於一區墙之伽%丨赞 、、 尾之個別苐二匯流排片段耦合至該全 域k擇匯机排,亚在特定其他時候耗合至該全域偏壓 線’並在其料候使個料二匯制#段浮動; 在該第-操作模式下,在該全域偏壓線上傳遞一適合 於”亥第#作換式之未選定位元偏壓電慶;以及 在該第二操作模式下,在該全域偏I線上傳遞-適合 於該第二操作模式$去、西h , 51 卜撰式之未選定位元偏壓電壓。 •一種用於製造一記纟 G體產品之方法,該方法包含·· 形成一記憶體陣列,复且 其具有複數個陣列區塊,各陣列 區塊包含字線及位元線,· 形成一第一全域匯流排,苴一 塊,用於有時將-選定區/、 複數個陣列區 料電路; 土口刃貝 形成每陣列區塊的_個 後^ ^ ^ ⑺弟匯机排片段,其用於在 第一刼作模式期間將一摘 f適合於該第一操作模式之第一 123178.doc -12- 200826114 線偏μ條件傳遞至―選^區塊之未選定位元 偏麼侔Γ適合於該第—操作模式之第二未選定位元線 5m料至未選定㈣區塊之未較位元線。 如鲕求項51之方法,其中: °亥第一操作模式包含一設定操作模式; ^於該第-操作模式之㈣—未選定位元線偏職 幵匕3 一未選定位元線電壓;以及 ^㈣第-操作模式之該第:未選定位元線偏麼條 件包含一浮動條件。 53·如請求項52之方法,其中·· :-第二操作模式下’一選定陣列區塊之該個別第一 =排片段絲合以將適合於㈣二操作模式之在個別 弟一匯流排片段匯流排線上的個別偏壓條件傳遞至該選 定區塊之選定位元線。 54. 如請求項53之方法,其進一步包含: 形成該第-全域匯流排作為一輕合至一讀取/寫入電路 之專用讀取/寫入匯流排。 55. 如請求項54之方法,其進一步包含: 形成一麵合至一第二資料電路之第二全域匯流排; 針對各區塊’形成一個別搞合電路’其用於有時將該 個別第-匯流排片段耦合至該第二全域匯流排,有時使 之浮動。 5 6 ·如凊求項5 3之方法,其進一步包含: 形成一耦合至一偏壓電路之全域偏壓線; 123178.doc -13 - 200826114 針對各區塊,形成一個別耦合電路,其用於有時將該 個別第一匯流排片段之各匯流排線耦合至該全域偏壓 線’在此期間該全域偏壓線傳遞一適合於該第一操作模 式之未選定位元線偏壓電壓。 57.如請求項56之方法,其進一步包含: 形成複數個耦合電路,其用於有時將用於各個別陣列 區塊之該等個別第一匯流排片段耦合在一起,以形成跨 越該複數個陣列區塊之一單一邏輯匯流排,該等耦合在 之的弟匯流排片段係麵合至該第二資料電路。 5 8·如請求項56之方法,其進一步包含: 針對各區塊,形成一個別耦合電路,其用於有時將該 個別第一匯流排片段耦合至該第二全域匯流排。 5 9 ·如請求項5 6之方法,其中·· 該第一全域匯流排包含耦合至該等個別資料電路之一 全域選擇匯流排;以及5. The method of claim 48, further comprising: coupling the gamma-supplied, and singular-single-bus two-segment segments for a zone wall to the global k-selection machine row, At other times, it is consumed by the global bias line 'and floats in the material of the second material system. In this first operation mode, a suitable one is transmitted on the global bias line. Unselected locating element bias voltage; and in the second mode of operation, transmitting on the global omni-directional I line - suitable for the second mode of operation $ go, west h, 51 puzzling unselected locating element Bias voltage. A method for fabricating a G-body product, the method comprising: forming a memory array having a plurality of array blocks, each array block comprising word lines and bit lines , forming a first global bus, one piece, for sometimes - selected area /, multiple array area material circuit; earth mouth blade forming each array block _ post ^ ^ ^ (7) Diocese a row segment for adapting a pick f during the first mode of operation The first 123178.doc -12- 200826114 of the operation mode is passed to the unselected positioning element of the "selection block". The second unselected positioning element line suitable for the first operation mode is 5m. The method of claim 51, wherein: the first operation mode includes a set operation mode; (in the first operation mode (4) - the unselected position line is not selected. Partial job 3 unselected location line voltage; and ^ (4) the first mode of operation - the unselected location line condition includes a floating condition. 53. The method of claim 52, wherein: - in the second mode of operation, the individual first = row segments of a selected array block are wired to pass individual bias conditions on the individual bus-bus segment bus lines suitable for the (four) two mode of operation to the selected The block selection location line 54. The method of claim 53, further comprising: forming the first-global bus as a dedicated read/write bus that is coupled to a read/write circuit. 55. The method of claim 54, further comprising Forming a second global busbar that is coupled to a second data circuit; forming a separate circuit for each block for coupling the individual first bus bar segment to the second global bus The method of claim 5, wherein the method further comprises: forming a global bias line coupled to a bias circuit; 123178.doc -13 - 200826114 for each block, Forming an additional coupling circuit for sometimes coupling respective bus bars of the respective first bus bar segments to the global bias line ' during which the global bias line is transmitted for a first mode of operation The positioning element line bias voltage is not selected. 57. The method of claim 56, further comprising: forming a plurality of coupling circuits for coupling together the individual first busbar segments for respective array blocks to form across the complex number One of the array blocks is a single logic bus, and the coupled bus segments are coupled to the second data circuit. The method of claim 56, further comprising: forming, for each of the blocks, a separate coupling circuit for coupling the individual first busbar segments to the second global busbar. 5. The method of claim 5, wherein the first global bus includes a global selection bus coupled to one of the individual data circuits; 广*進-步形成每陣列區塊的一個別第二匯流排片 段’其用於在該第一操作模式期間,將一適合於該第一 插作模式之選定位元線㈣條件傳遞至m塊之選 =元線,並料在該第二操作模式期間,將—適合於 =操作模式之未選定位元線偏堡條件傳遞至未選定 陣列區塊之未選定位元線。 60.如請求項51之方法,其進一步包含·· 元:成-三維記憶體陣列’其具有在二位元線層上的位 123178.doc -14- 200826114 61.如請求項60之方法,其中: 電阻器元件。 形成記憶體單元,其包含一可逆Extendingly forming a second bus bar segment of each array block for transmitting a conditional element (4) condition suitable for the first interpolation mode to the m during the first mode of operation The block selection = element line, and during the second mode of operation, the unselected positioning element line conditions suitable for the = operating mode are passed to the unselected positioning elements of the unselected array block. 60. The method of claim 51, further comprising: a meta-three-dimensional memory array having a bit on a two-bit line layer 123178.doc -14 - 200826114 61. The method of claim 60, Where: Resistor component. Forming a memory unit that includes a reversible 123178.doc 15-123178.doc 15-
TW96128078A 2006-07-31 2007-07-31 Method and apparatus for hierarchical bit line bias bus for block selectable memory array TWI345787B (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US11/461,376 US7596050B2 (en) 2006-07-31 2006-07-31 Method for using a hierarchical bit line bias bus for block selectable memory array
US11/461,362 US7633828B2 (en) 2006-07-31 2006-07-31 Hierarchical bit line bias bus for block selectable memory array

Publications (2)

Publication Number Publication Date
TW200826114A true TW200826114A (en) 2008-06-16
TWI345787B TWI345787B (en) 2011-07-21

Family

ID=38997824

Family Applications (1)

Application Number Title Priority Date Filing Date
TW96128078A TWI345787B (en) 2006-07-31 2007-07-31 Method and apparatus for hierarchical bit line bias bus for block selectable memory array

Country Status (2)

Country Link
TW (1) TWI345787B (en)
WO (1) WO2008016951A2 (en)

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5245570A (en) * 1990-12-21 1993-09-14 Intel Corporation Floating gate non-volatile memory blocks and select transistors

Also Published As

Publication number Publication date
WO2008016951A3 (en) 2008-05-15
WO2008016951A2 (en) 2008-02-07
TWI345787B (en) 2011-07-21

Similar Documents

Publication Publication Date Title
US7554832B2 (en) Passive element memory array incorporating reversible polarity word line and bit line decoders
US7463546B2 (en) Method for using a passive element memory array incorporating reversible polarity word line and bit line decoders
US7486587B2 (en) Dual data-dependent busses for coupling read/write circuits to a memory array
US7499366B2 (en) Method for using dual data-dependent busses for coupling read/write circuits to a memory array
US8913413B2 (en) Memory system with sectional data lines
US7570523B2 (en) Method for using two data busses for memory array block selection
US7463536B2 (en) Memory array incorporating two data busses for memory array block selection
US7633828B2 (en) Hierarchical bit line bias bus for block selectable memory array
US7596050B2 (en) Method for using a hierarchical bit line bias bus for block selectable memory array
TW201214460A (en) Decoder circuitry providing forward and reverse modes of memory array operation and method for biasing same
KR101478193B1 (en) Method and apparatus for passive element memory array incorporating reversible polarity word line and bit line decoders
TWI345790B (en) Method and apparatus for dual data-dependent busses for coupling read/write circuits to a memory array
TWI345791B (en) Method and apparatus for memory array incorporating two data busses for memory array block selection
TW200826114A (en) Method and apparatus for hierarchical bit line BIAS bus for block selectable memory array

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees