TW201133504A - Current sink system based on sample and hold for source side sensing - Google Patents

Current sink system based on sample and hold for source side sensing Download PDF

Info

Publication number
TW201133504A
TW201133504A TW99108947A TW99108947A TW201133504A TW 201133504 A TW201133504 A TW 201133504A TW 99108947 A TW99108947 A TW 99108947A TW 99108947 A TW99108947 A TW 99108947A TW 201133504 A TW201133504 A TW 201133504A
Authority
TW
Taiwan
Prior art keywords
current
transistor
voltage
memory cell
time interval
Prior art date
Application number
TW99108947A
Other languages
Chinese (zh)
Other versions
TWI442408B (en
Inventor
Kuo-Yu Liao
Han-Sung Chen
Chun-Hsiung Hung
Original Assignee
Macronix Int Co Ltd
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
Application filed by Macronix Int Co Ltd filed Critical Macronix Int Co Ltd
Priority to TW99108947A priority Critical patent/TWI442408B/en
Publication of TW201133504A publication Critical patent/TW201133504A/en
Application granted granted Critical
Publication of TWI442408B publication Critical patent/TWI442408B/en

Links

Landscapes

  • Read Only Memory (AREA)

Abstract

Source side sensing techniques described herein determine the data value stored in a memory cell based on the difference in current between the read current from the source terminal of the memory cell and a sink current drawn from the read current. The sink current is drawn in response to the magnitude of an operating voltage between first and second nodes. During a first time interval, the operating voltage is set in response to a magnitude of the reference current using a feedback path. During a second time interval following the first. time interval, the operating voltage is held independent of the feedback path. The data value stored in the memory cell is determined based on a difference in current between the read current and a sink current during the second time interval.

Description

201133504 六、發明說明: 【發明所屬之技術領域】 關於記憶裝置的資料感測,特別是關於如此防止記 憶裝置中因為雜訊所造成的感測操作干擾。 【先前技術】 現今存在料_電荷儲存賴胞型態轉揮發記憶體 括,憶胞儲存電荷於-場效電晶體的通道與_之間。所儲 電荷數里了電晶體的臨界賴,其可以被感取指示資料。 二種型態的電荷儲存記憶胞被稱為浮動閘極記憶胞。在浮 =記憶胞巾’ f荷被儲存在-電性層介於場效電晶體的通 道與閘極之間。臨界電壓的改變藉由施加一合適的電壓於 層儲存或移除電荷。另一種型態的記憶胞被稱為 電何捕巧5己憶胞,其使用一介電電荷捕捉層取代浮動閘極。 、在讀,或感測儲存在一記憶胞中的資料值之操作時,係施加 合適的電壓以自此記憶胞的汲極端誘發電流至源極端。此電流 取决於電晶體的臨界電壓,因此指示儲存於其中的資料。 “ —決定儲存在一記憶胞中的資料值可以利用一感測放大器來進 行’其感測流經此記憶胞的電流,且將此感測的電流與一合適的 值作比較。第1圖顯示一習知技術的感測放大器17〇的實施 不意圖,其可以感測儲存在一選取記憶胞110中的資料值。 s己憶胞1丨〇是一記憶陣列中的代表性記憶胞,此記憶陣列包 1數百萬或數十億個記憶胞。字元線12〇與記憶胞11〇的閘極終 端耦接。位元線13〇、132與記憶胞11〇的源極和汲極終端112'、、 114耦接。行選擇電晶體140係響應一 SEL信號以選擇性地將位 =線130與連接至感測放大器170感測輸入172的資料線15〇耦 在δ己憶胞110的感測操作中,係施加合適的電壓至字元線 及位元線丨32誘發一讀取電流Icell自汲極終端U4至源極 終端U2及進入位元線130。此讀取電流ICELL經由行選擇電晶 201133504 體140被提供至資料線15〇。 17 0之感測輸入17 2來對一等㈣都雷六,社?= 3 π。寻政員戟電谷CL0A⑴充電,導致感 ^t s 坚改變係與在此讀取操作時間中的讀取電流 =ELL目關。^項取電流ICELL的大小係與記憶胞u〇的臨界電 壓相關’具有較低的臨界電壓可導致—較大的電流。因此,感 測輸入17 ^的電壓會改變的較快,假如此記憶胞i i G在一較低 的臨界狀態而不是在一較高的臨界狀態。 ” ms作中,使用一參考電流1咖提供至在感測放大 窃170之感測輸入172以作為參考值。此例示的範中,來 胞210用來提供參考電流Iref。 」201133504 VI. Description of the Invention: [Technical Field of the Invention] The data sensing of the memory device is particularly concerned with the prevention of interference in the sensing operation caused by noise in the memory device. [Prior Art] There is a material stored in the charge-type memory to the volatile memory, and the memory is stored between the channel and the _ of the field effect transistor. The number of stored charges is the critical value of the transistor, which can be sensed. Two types of charge storage memory cells are called floating gate memory cells. In the floating = memory cell, the f load is stored between the channel and the gate of the field effect transistor. The change in threshold voltage stores or removes charge at a layer by applying a suitable voltage. Another type of memory cell is called an electron cell, which uses a dielectric charge trapping layer instead of a floating gate. When reading, or sensing the operation of storing data values in a memory cell, a suitable voltage is applied to induce current from the 汲 terminal of the memory cell to the source terminal. This current depends on the threshold voltage of the transistor and therefore indicates the data stored in it. “—Determining the value of the data stored in a memory cell can be performed using a sense amplifier that senses the current flowing through the memory cell and compares the sensed current to an appropriate value. Figure 1 The implementation of a conventional technique of the sense amplifier 17A is not intended to sense the value of a data stored in a selected memory cell 110. The memory cell is a representative memory cell in a memory array. The memory array package has millions or billions of memory cells. The word line 12〇 is coupled to the gate terminal of the memory cell 11〇. The bit lines 13〇, 132 and the source and the channel of the memory cell 11〇 The pole terminals 112', 114 are coupled. The row select transistor 140 is responsive to a SEL signal to selectively couple the bit = line 130 to the data line 15 coupled to the sense amplifier 170 sense input 172. In the sensing operation of the cell 110, applying a suitable voltage to the word line and the bit line 丨32 induces a read current Icell from the drain terminal U4 to the source terminal U2 and into the bit line 130. This read current ICELL is supplied to the data line 15 via the row selection cell 201133504 body 140. 17 0 sense transmission Into 17 2 to the first class (four) Dulei six, the community? = 3 π. The traitor 戟 谷 谷 CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL CL The magnitude of the current ICELL is related to the threshold voltage of the memory cell u〇. A lower threshold voltage can cause a larger current. Therefore, the voltage of the sensing input 17 ^ will change faster, if so The memory cell ii G is in a lower critical state rather than a higher critical state. In the ms, a reference current 1 is used to provide a sense input 172 at the sense amplification 170 as a reference value. In the illustrated example, the cell 210 is used to provide a reference current Iref. "

字元線220與參考胞21〇的閘極終端耦接。位元線23〇、 232與參考胞210的源極和汲極終端2丨2、214耦接。 電晶體142係、選擇性地將位元線23〇肖連接至感測放大器17〇 參考輸入174的參考線16〇耦接。 合適的電壓被施加至字元線220及位元線232此記憶胞以 誘發一參考電流Iref自汲極終端214至源極終端212及進入位 元線230。此參考電流IREF經由行選擇電晶體M2被提供至參 考線160。此參考電流IREF在此感測放大器17〇之參考輸入口4 來對一等效負載電容cLOAD2充電,導致參考輸入174的電壓 轉換此參考電流Iref為一參考電壓。 感測放大器170在感測操作時係由感測致能信號sen所 觸發。此感測放大器170係響應感測輸入172與參考輸入174 之間的電壓差值,而產生用來指示儲存於該選取記憶胞11〇 的一輸出信號176。 ~ 第2圖為在記憶胞110在感測操作時感測輸入172與夂 輸入174的電壓改變與時間的關係簡要示意圖。曲線25〇'^示 假如§己憶胞1 ίο是在一較低臨界狀態時感測輸入的電壓改 變’而曲線260顯示假如S己憶胞110是在一較高臨界狀痒 測輸入172的電壓改變。曲線250和260之間的差異g時^ T1時是用以區分此記憶胞是在一較低還是較高臨界^態的^ 201133504 ίί J2,區間。為了可#地區分較高與較低臨界狀態,必 直'且;憶胞的實施例中, 其具有超過兩個臨界狀態,其間並具有感測區間。 严”,Ϊ線Γ〇顯示記憶胞110在感測操作時參考輸入174的電 3 在此例示中,曲線27G*時間T1具有—介於較J 界狀,曲線250與較高臨界狀態曲線26〇之間的一電壓。如 考胞21G的臨界電壓介於記憶胞 右八較*臨界狀態之間而達成’如此參考電流I卿且 有"於汜憶胞no較低與較高臨界狀態的讀取電流 ^了大小。而在另一範例中,如此可以藉由施加不同的電壓丄 子兀線120和220 ’及/或施加不同的電壓至位元線132和23 而達成。 此感測放大器170產生一輸出信號176,其具有一數值曰 根據在時間T1時感測輸入172的電壓是大於或小於參考輸= 174的電壓而定,因此指示儲存於該選取記憶胞丨中的資 値0 、’ 此感測過程中會因為感測放大器17〇容易受到雜訊的影 響而產生問題。特別是,於感測操作中的雜訊會影響感測輸入 Π2與參考輸入174之間的電壓差值,其會增加感測放大器17〇 的複雜程度或是感測所需的時間。 在第1圖所示的範例中,記憶胞110是與參考胞21〇隔 離,且此讀取電流ICELL與參考電流Iref並不會彼此依賴。其 結果是,記憶胞110可以暴露在與參考胞21〇不同的雜訊下二 其可以導致讀取電流ICELL與參考電流Iref不同比例的改變。 其結果是一個較廣的感測輸入Π2與參考輸入174之間的電壓 差值變動,其抑制了感測放大器17〇可以正確地感測儲存於該 選取記憶胞110中的資料値。 在上述的實施例中’感測放大器170的感測輸入172是與 。己丨思胞110的源極終端112轉接(源極端感測)。其結果是,源 極終端112上的電壓也會增加一個與讀取電流IcEu^a關的數 201133504 量。此,極終端112上的電壓增加減少了及極至源極 效應’其因此會降低了由記憶“ 因為操作的環境變動的關係,也會因為 ===記;f胞的臨界電壓會在陣列⑽之間變動tti ”造,陣列1〇5間儲存相同資料時讀取電流w c l括因為源極電壓增加所導致的讀取電流认改變 丨=電壓增加的一數量與讀取電流W相關,1 加了感應感測電路的複雜程度或是所需的: 自資料線輕—祉_祕端 =問=之 因為源極電產及且方古法’其可以解決在操作時 ^ η級痛所產生的問題’及具有對雜訊的較高承受能力。 【發明内容】 所;,=源極端感測技術根據自此記憶胞源極終端 ΐΐίίί:該讀取電流所導入之汲入電流之間的差Ϊ ίΐΐΐίίΞ憶胞中的資料值。此汲入電流係響應由-介於 ί丄;二的操作電壓大小而導入。於第-時間區間 設定介Α杨㈣-晴職射以快速地 用一電容琴以間間中失該回授路徑且該操作電壓利 的。第λ賴區财基本上是定值 區間中基本上貞似的n如此操作電_二時間 取故胞中的資料值是根據在第二_區財 201133504 不會被= 定。因此,此差值的正確性基本上 使用讀取ΐ流 流,會減少在感測操作時的差值,而不是整個讀取電 會在此陣财的記#極終_電壓魏。如此則 的變動。其結果是,於此;原極端感測時減少讀取電流 電流分佈會變得較為緊縮。Μδ2ί憶胞中的感測節點之電壓或 此處所描述之記情梦罢4 k 中的-選取記憶胞提供一^^ 記憶陣列可自該記憶陣列 可提供-參考電流,」—資料線’—參考電流源 電流源可響應介於該第:/第在輕|妾的沒入電流源,該沒入 小而自該資料線導入ΐ及入電::3間作電壓的一大 ί 1^ ;:r; 門以^〜減電壓’―電容器输在該第—與第—節^之 二:月::亥回授路徑,及在該第一時間區間之後 】 i間以及—感測放大電路響應於該第二時間 干電流與該沒人電流之間的差值,而產生-用來指 ”儲,於賴選取記憶胞中的資料值之輸出信號。 胞以自記憶胞的方法包含施加一偏壓至該記憶 的—楚與第一郎點間的_作電壓,及在該第一時間區間之後 益關中Γϊί於該第—與第二節點間與該回授路徑 作電壓,根據在該第二時間區間中該讀取電流與該汲 電机之間的一差值決定儲存於該記憶胞中的一資料值。 節中述特徵’和實施例,會在下列實施方式的章 201133504 【實施方式】 本發明實施例搭配以下第3到9圖進行詳細描述。 第3圖顯示一積體電路3〇〇的簡要方塊示意圖,在 含本發明所描述之使用於—記憶陣列32〇中的取樣和保上 汲入電流電路310’其可以用來感測儲存在記憶陣 〇、^ 胞中的資料值。 °己隐 —列解碼器322與複數條沿著此記憶陣列32〇列方向上 mt324耦接。行解碼器326與複數條沿著此記憶^ 3 〇仃方向上排列的位元線328耦接,以自此陣列的 感:二Ϊ式,抹除。此記憶陣列320中的記憶胞可 , 、平行地或是虛擬接地的方式排列。在區塊330中的 結構在此範例中經由資料匯流㈣ 旳、隹二!^下更詳細地描述,對此記憶陣列320的一選取記憶 ”^__時’適當的電顧施加 | 32G的—選取記憶胞之源極端^ 響應介於if·!之,之沒人電流電路31G與資料線332耦接以 又二汲入雷、、* Γ與第二節點之間的—操作電壓的大小而導 介於此取樣i伴此處所使用的名詞,,操作電壓,,通常是指一 點之間的及入電流電路3ig的第一節點與第二節 變的話,而建立:具料了響應參考電流1咖的大小若是改 以下所ίΐϊ—改變沒入電流1·大小的電壓。 授路徑可以和㈣之汲人電流電路训,包括-回 一節點與第二應參考電流1REF的大小而操作設定介於第 區間(取樣區間^巾之間的操作電壓。於感測操作的一第一時間 壓。於感測捍竹仏,此回授路徑被致能以很快地設定此操作電 被失能,且一、俾姓:第二時間區間(保持區間)中,此回授路徑 保持此操作電^ #電容器被耦接在第一節點與第二節點之間以 I乎疋一個定值且因此大致上是沒有雜訊的。 201133504 在此處所描述的實施例中,此操作電愿 持之汲入電流電路3H) t的負載電晶體控 端=:導通終端(例如沒極或源極終端)之= -Dr Ϊ參考電流Iref。此保持電容器是用來施加-大致 的控制-導通終端電壓(例如閘極 = 欠 所是=細的電流在此保持心^ 擁33^2電Ϊ 1眶和參考電流“之間的差值是提供區 感ΪΪ大器電路。此感測放大器電路係響應第二時 才曰不儲存在所選取記憶胞中的一資料值的輸出信號。 王 i 中所不,此取樣和保持之沒入電流電路310也會 t’_-參考_ 34G t參考胞所產生的參考 冤〜IREFk供給取樣和保持之汲入電流電路31〇。替代地也可 以使用其他的技術來產生參考電流IREF。舉例而言,參考電流 REF可以藉由根據超過一個以上的參考胞之參考電流。 列解碼器344與沿著此參考陣列34〇列方向上排列的字元 巧345耦接。行解碼器342與沿著此參考陣列34〇行方向上排 =的位兀線343耦接。在此例示實施例中,參考陣列34〇與記 ^車列320是分開的,且包括分別的列及行解碼器344、342°。 f代地,參考陣列340可以是在記憶陣列32Θ中的一部分,且 在陣列320、340中分享解碼器。 。。位址係透過匯流排35〇提供至行解碼器326、342及列解 =322、344。資料係由積體電路3〇〇上的輸入/輸出埠透過 輸入線352傳送至方塊330之資料輸入結構。在此例示的 例中,其他電路360也包括在此積體電路3〇〇内,例如通 八目的處理器或特殊用途電路’或是由此記憶陣列所支援的組 5模組以提供單晶片系統功能。資料係由方塊330中的感測放 201133504 大器,透過資料輸出線354,傳送至積體電路300上的輸入/ 輸出埠或其他積體電路300内或外之資料目的地。 此積體電路300包含控制器369以感測、程式化和抹除記 憶陣j 320記憶胞與參考陣列34〇中的參考胞。此控制器369, 在,範例中為一偏壓調整狀態機構,控制由區塊368中產生的 偏壓調整供應電壓或提供至區塊368中的讀取、程式化或抹除 電壓。此控巧器369的應用可以使用,業界所熟知的技術,如 特殊目的邏輯電路來實施。在另一實施例中,該控制器369包 含一通用目的處理器,其可以實施在相同積體電路上,其執行The word line 220 is coupled to the gate terminal of the reference cell 21A. The bit lines 23A, 232 are coupled to the source and drain terminals 2A2, 214 of the reference cell 210. The transistor 142 is coupled to selectively connect the bit line 23 to the sense amplifier 17 参考 the reference line 16 of the reference input 174. A suitable voltage is applied to word line 220 and bit line 232 to induce a reference current Iref from drain terminal 214 to source terminal 212 and to bit line 230. This reference current IREF is supplied to the reference line 160 via the row selection transistor M2. The reference current IREF charges the equivalent load capacitance cLOAD2 at the reference input port 4 of the sense amplifier 17A, causing the voltage of the reference input 174 to convert the reference current Iref to a reference voltage. The sense amplifier 170 is triggered by the sense enable signal sen during the sensing operation. The sense amplifier 170 is responsive to a voltage difference between the sense input 172 and the reference input 174 to generate an output signal 176 for storage in the selected memory cell 11A. ~ Fig. 2 is a schematic diagram showing the relationship between the voltage change of the sense input 172 and the 夂 input 174 and the time when the memory cell 110 is in the sensing operation. The curve 25 〇 ' 示 假 § § 忆 忆 忆 1 ί 忆 忆 忆 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感 感The voltage changes. The difference between curves 250 and 260, when T1 is used to distinguish whether the memory cell is in a lower or higher critical state, is 201133504 ίί J2, the interval. In order to distinguish between higher and lower critical states, it is necessary to have more than two critical states in the embodiment of the cell, with a sensing interval therebetween. Strictly, the line Γ〇 shows that the memory cell 110 refers to the power of the input 174 during the sensing operation. In this illustration, the curve 27G* time T1 has - between the J boundary, the curve 250 and the higher critical state curve 26 A voltage between the 〇, such as the threshold voltage of the 21G of the cell is between the right octet and the *critical state of the memory cell to achieve 'such reference current I qing and has " Yu 汜 胞 no low and higher critical state The read current is sized. In another example, this can be achieved by applying different voltages to the turns 120 and 220' and/or applying different voltages to the bit lines 132 and 23. The sense amplifier 170 produces an output signal 176 having a value 而 depending on the voltage at which the sense input 172 is greater or less than the reference input = 174 at time T1, thus indicating the amount of money stored in the selected memory cell.値0, ' This sensing process may cause problems due to the influence of the sense amplifier 17〇. In particular, the noise in the sensing operation affects the sense input Π2 and the reference input 174. Voltage difference, which increases the sense amplifier 17 The complexity or the time required for sensing. In the example shown in Figure 1, the memory cell 110 is isolated from the reference cell 21〇, and the read current ICELL and the reference current Iref are not dependent on each other. As a result, the memory cell 110 can be exposed to a different noise than the reference cell 21, which can result in a change in the read current ICELL from the reference current Iref. The result is a wider sense input Π2 and reference input. The voltage difference between 174 varies, which inhibits the sense amplifier 17A from correctly sensing the data stored in the selected memory cell 110. In the above-described embodiment, the sense input 172 of the sense amplifier 170 The source terminal 112 of the signal cell 110 is switched (source terminal sensing). As a result, the voltage on the source terminal 112 is also increased by a number 201133504 that is off from the read current IcEu^a. Thus, the voltage increase on the pole terminal 112 is reduced and the pole-to-source effect 'which is therefore reduced by the memory" because of the environmental variation of the operation, and also because of ===; the threshold voltage of the f cell will be in the array (10) change between tti" When the array 1〇5 stores the same data, the read current wcl includes the read current change caused by the increase of the source voltage. 丨=the amount of voltage increase is related to the read current W, and the inductive sensing circuit is added. Complexity or required: Since the data line is light - 祉 _ secret end = Q = because the source of electricity and the ancient method 'which can solve the problem of η grade pain during operation' and have The invention has the high tolerance of the noise. [Invention] The source-source sensing technology is based on the memory source terminal ΐΐίίί: the difference between the inrush currents introduced by the read current Ϊ ΐΐΐ Ξ Ξ Ξ Data value. This inrush current is introduced in response to the operating voltage of -. In the first time interval, set the Yang (4)-Qing yin to quickly use a capacitor to lose the feedback path between the two and the operating voltage is favorable. The λ 赖 区域 财 is basically a fixed value in the interval is basically similar to the n operation _ two time to take the data value in the cell is based on the second _ district 201133504 will not be = fixed. Therefore, the correctness of this difference basically uses the read turbulent flow, which reduces the difference in the sensing operation, rather than the entire reading power. This is the change. As a result, this reduces the read current and the current distribution becomes tighter during the original extreme sensing. Μ δ2 ί ί 的 ί ί 或 或 或 或 或 或 或 或 ί ί ί ί ί ί ί ί ί ί ί ί 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取 选取The reference current source current source can respond to the immersed current source between the first:/the first light|妾, which is small and is introduced from the data line and the incoming power: 3 is a large voltage ί 1^; :r; The gate is ^~voltage-reducing--capacitor is in the first-and-th-section^2: month::Hai feedback path, and after the first time interval] i and - sense amplifier circuit In response to the difference between the second time dry current and the unmanned current, the output signal is used to refer to the data value of the selected memory cell. The cell contains the self-memory cell method. a voltage is applied to the voltage between the first and second points of the memory, and after the first time interval, the voltage is applied between the first and the second node and the feedback path, according to a difference between the read current and the 汲 motor in the second time interval is determined to be stored in the memory A data value in the section. The features and embodiments in the section will be in the following embodiment of the chapter 201133504. [Embodiment] The embodiment of the present invention is described in detail with the following figures 3 to 9. FIG. 3 shows an integrated circuit. BRIEF DESCRIPTION OF THE DRAWINGS A schematic block diagram of a sample and hold-in current circuit 310' for use in a memory array 32A as described herein can be used to sense storage in memory cells and cells. The data value is coupled to the mt 324 in the direction of the memory array 32. The row decoder 326 and the plurality of bit lines arranged along the memory ^ 3 〇仃 direction The 328 is coupled to the sense of the array: the eraser. The memory cells in the memory array 320 can be arranged in parallel, or in a virtual ground. The structure in the block 330 is in this example. According to the data convergence (4) 旳, 隹二! ^, in more detail, a memory of the memory array 320 is selected "^__" when the appropriate application is applied | 32G - the source of the selected memory cell ^ response If·!, no one current circuit 31G and data 332 is coupled to the magnitude of the operating voltage between the second, the *, and the second node, and is used to refer to the noun used herein, the operating voltage, and usually refers to a point between And the first node of the current input circuit 3ig and the second node change, and establish: the magnitude of the response reference current 1 coffee is changed according to the following - change the voltage of the inrush current 1 · size. The routing path can be operated with (4) the current circuit, including the back-to-one node and the second reference current 1REF, and the operating voltage is set between the sampling interval (the sampling interval). The first time pressure. In the sense of the bamboo pole, this feedback path is enabled to quickly set the operation to be disabled, and one, the last name: the second time interval (hold interval), this feedback The path maintains this operation. The capacitor is coupled between the first node and the second node to have a constant value and is therefore substantially free of noise. 201133504 In the embodiment described herein, this operation The load transistor terminal of the current input circuit 3H) t =: the conduction terminal (for example, the pole or the source terminal) = -Dr Ϊ reference current Iref. This holding capacitor is used to apply - roughly control - turn-on terminal voltage (eg gate = under-react = fine current here keeps the heart ^ 33 ^ 2 electric Ϊ 1 眶 and reference current " the difference between The sense amplifier circuit is provided. The sense amplifier circuit responds to the output signal of a data value in the selected memory cell in response to the second time. In the case of the i, the sampled and held immersed current The circuit 310 also supplies the reference 冤~IREFk generated by the reference _ 34G t reference cell to the sampled and held mash current circuit 31. Alternatively, other techniques can be used to generate the reference current IREF. The reference current REF may be based on a reference current based on more than one reference cell. The column decoder 344 is coupled to a word 345 arranged in the direction of the reference array 34. The row decoder 342 is along The reference array 34 is coupled to the bit line 343 in the row direction. In this exemplary embodiment, the reference array 34 is separated from the row of columns 320 and includes separate column and row decoders 344, 342°. f generation, the reference array 340 can be A portion of the memory array 32A, and the decoders are shared among the arrays 320, 340. The address is provided to the row decoders 326, 342 and the column solutions = 322, 344 through the bus bar 35. The data is integrated. The input/output port on circuit 3 is transmitted through input line 352 to the data input structure of block 330. In the illustrated example, other circuits 360 are also included in the integrated circuit 3, for example, eight-way processing. Or a special-purpose circuit' or a group 5 module supported by the memory array to provide a single-chip system function. The data is transmitted from the sense output in block 330 to the integrated body through the data output line 354. Input/output on circuit 300 or other data destinations within or outside of integrated circuit 300. This integrated circuit 300 includes controller 369 to sense, program, and erase memory array 320 memory cells and reference array 34 The reference cell in the cymbal. The controller 369, in the example, is a bias adjustment state mechanism that controls the bias voltage generated by the block 368 to adjust the supply voltage or provide a read, stylized or block to block 368. Wipe off the voltage. This control The application of the device 369 can be implemented using techniques well known in the art, such as special purpose logic circuits. In another embodiment, the controller 369 includes a general purpose processor that can be implemented on the same integrated circuit. carried out

一,腦程式以控制該裝置的操作。在另一實施例中,特殊目的 邏輯電路和一通用目的處理器的組合可以被用來實施該控 器 369 〇 第4圖顯示此記憶陣列細中的—選取記憶胞進行感測操 作之一方法400流程圖,其可以由控制器369來執行。 在方塊410,施加一讀取偏壓至該被選取之記憶胞以自該 被選取之記憶胞的-源極終端誘發—讀取電流w至資料以 332 ° 在万塊,自一參考電流源提供-參考電流IREF。在第 3圖的H實施例中,此參考電流^是料施加合 電壓至參考陣列340中的參考胞而提供。 碰 如=之前所描述的,自資料線332 汲 係響應^於此,,持线入電流電路31〇(方塊; 的第-即點與第-即點之間的操作電壓大 間(方塊例,取樣和保持线人電流電路3m 徑可以用來響應參考電流lREF的大 : 點與第二節點=的操作電壓。將於以下更詳細=述第此= 授路徑可以很快地^此取樣和保持线人電流電路的 ^占,包括很快地設定介於第—節點與第二節點之間的操作 壓。 201133504 於第一時間區間後的一第二時間區間(方塊440)中,此操 作電壓被保持在第一節點與第二節點之間而與回授路徑無關。 在方塊450 ’根據讀取電流ICELL與該汲入電流iSINK之間 在第一時間區間中的一差值決定儲存於該被選取記憶胞中的 資料值。 第5圖顯示本發明第一實施例之使用於一記憶陣列320中 的取樣和保持之汲入電流電路31〇的簡要示意圖,其可以用來 感測儲存在記憶陣列320所選取記憶胞510中的資料值。 字元線324a與所選取記憶胞510的閘極耦接。位元線 328a、328b分別與記憶胞510的汲極終端511和源極終端512First, a brain program to control the operation of the device. In another embodiment, a combination of a special purpose logic circuit and a general purpose processor can be used to implement the controller 369. Figure 4 shows a method of selecting a memory cell for sensing operation in the memory array. 400 flowchart, which may be executed by controller 369. At block 410, a read bias is applied to the selected memory cell to induce from the source terminal of the selected memory cell - the read current w to the data is 332 ° in 10,000 volts from a reference current source Provide - reference current IREF. In the H embodiment of Figure 3, this reference current is provided by applying a combined voltage to a reference cell in reference array 340. As described earlier, since the data line 332 is responsive to this, the line-in current circuit 31 is clamped (block; the operating voltage between the first-point and the-point-point is large (block example, The sample and hold line current circuit 3m diameter can be used to respond to the large reference voltage lREF: the operating voltage of the point and the second node = will be described in more detail below = the current = the path can be quickly sampled and maintained The occupation of the line current circuit includes quickly setting the operating pressure between the first node and the second node. 201133504 This operating voltage is in a second time interval (block 440) after the first time interval. Is maintained between the first node and the second node regardless of the feedback path. At block 450', a difference between the read current ICELL and the inrush current iSINK in the first time interval is determined to be stored in the The data values in the selected memory cells. Fig. 5 is a schematic diagram showing the sampling and holding of the inrush current circuit 31A used in a memory array 320 according to the first embodiment of the present invention, which can be used to sense storage. Memory selected by memory array 320 The data value in cell 510. Word line 324a is coupled to the gate of selected memory cell 510. Bit line 328a, 328b is coupled to drain terminal 511 and source terminal 512 of memory cell 510, respectively.

摩禺接。 一在此記憶胞5丨〇的一感測操作時,適當的電壓被施加至字 元線324a和位元線328a以誘發一讀取電流Icell自汲極終端 511至源極終端512及進入位元線328b。此讀取電流iCEU經 由行解碼器326被提供至資料線332a。 在此例示的範例中,參考陣列340的參考胞560用來提供 參考電流IREF。字元線345a與參考胞560的閘極終端耦接。 位元線343a、343b分別與參考胞560的汲極終端561和源極 終端562搞接。Capricorn pick up. When a sensing operation of the memory cell 5 is performed, an appropriate voltage is applied to the word line 324a and the bit line 328a to induce a read current Icell from the drain terminal 511 to the source terminal 512 and the entry bit. Yuan line 328b. This read current iCEU is supplied to the data line 332a via the row decoder 326. In the illustrated example, reference cell 560 of reference array 340 is used to provide a reference current IREF. The word line 345a is coupled to the gate terminal of the reference cell 560. The bit lines 343a, 343b are respectively coupled to the drain terminal 561 and the source terminal 562 of the reference cell 560.

七電壓被施加至字元線345a和位元線343a以誘發一參考<5 流IrEF自汲極終端561至源極終端562及進入位元線3^^ 此參考電流iREF經由行解碼器342被提供至參考線346a。 此取樣和保持之沒入電流電路包括電晶體,其」 及極終端與參考線3恤耗接以接收參考電流丨。閉^ 知選擇性地經由切換開關526與參考線346a耦接。此例示^ 關52^使用單一通過電晶體具有一閘極耦接封 5虎S卜也可以使用其他的替代技來實 舉例而言,切換開關526可以使用一組互補通過具5^ 行的導通路径及提供互補的閘極控制信號。 八 12 201133504 和電ί S 3L526 電晶體522是二極體式連接以 和電日日體524構成一電流鏡520,電晶艚读 a 兩* , = 524 入5亥;及入電流iSINK,而此;:及人電流Isink具有—大小 電流Iref大小的一函數關係。舉例而言,汲入 厂_、'、二一 大小可以大致與參考電流IREF大小相當。 机SINK的一 在此例示的範例中,電流鏡52〇是由電晶體522、52 貫施。替代地也可以使用其他的技術來實施電流鏡52〇。 保持電容器C1 525是耦接介於電晶體522、524的 526 5Seven voltages are applied to word line 345a and bit line 343a to induce a reference <5 stream IrEF from drain terminal 561 to source terminal 562 and into bit line 3^^ this reference current iREF via row decoder 342 It is provided to reference line 346a. The sampled and held immersed current circuit includes a transistor, and the pole terminal is consuming with a reference line 3 shirt to receive a reference current 丨. The switch is selectively coupled to the reference line 346a via the changeover switch 526. The example switch 52^ uses a single pass transistor to have a gate coupling 5, and other alternative techniques can be used. For example, the switch 526 can be turned on using a set of complementary passes. The path and provide complementary gate control signals.八12 201133504 和电 S 3L526 transistor 522 is a diode connection to form a current mirror 520 with the electric solar 524, the electric crystal read a 2*, = 524 into 5 hai; and the current iSINK, and this ;: and the human current Isink has a function of the magnitude of the magnitude current Iref. For example, the size of the intrusion factory _, ', and two can be roughly equal to the magnitude of the reference current IREF. In the illustrated example, current mirror 52A is implemented by transistors 522, 52. Alternatively, other techniques can be used to implement the current mirror 52A. The holding capacitor C1 525 is coupled to the 526 5 of the transistors 522, 524.

保持電M Cl 525建立而維持在電晶體522、524 端。此保持電容H C1 525的大小足夠大可 _早The hold M Cl 525 is established and maintained at the ends of the transistors 522, 524. The size of this holding capacitor H C1 525 is large enough _ early

偏壓電壓維持在電晶體522、524的閘 終鈿之間,所以電晶體522、524保持開啟,而介於汲入電流 IsiNK與參考電流IREF之間的電流差值(假如有的話 L 時間區間保持大致是定值。 保符 "可以理解的是,由保持電容器C1 525將偏壓電壓維持在 電晶體522、524的閘極終端與源極終端之間在保持時間區間 或許會因為電荷洩漏以及切換開關526開啟的瞬間雜訊而& 變。 負載電晶體530具有一汲極終端與電晶體522、524的原 極終端耦接。負載電晶體530具有一源極終端與一負升壓器/電 路532耦接’其提供一個負的偏壓電壓。 、 ° 通過負載電晶體530的電流是在電晶體522的參考電汽 Iref與電晶體524的汲入電流ISINK之組合。通過負載電晶= 530的電流係由負載電晶體530將電晶體522、524的源:終 端之一電壓轉換而得。 、〜 此取樣和保持之j:及入電流電路31 〇也可以包括一運算放 大器540。供給電壓VDD和負升壓器電路532提供偏壓電壓 至此運算放大器540。在此例示中,此運算放大器54〇具有負 13 201133504 輸入544與地輕接。此運算放 線346a耦接。因為這兩於 ^ 一有正輸入540與參考 參考線346a上的參考電^ 和546虛擬接地的效應, 致為地電位。此名詞,,大f,,^EF ^算放大器540偏壓大 和546之間因為運算放 來表示介於兩輸入端544 成的-電屢差。.中的一非零輸入偏移電歷所造 負載地藉由切換開關57〇與 570是使用單一通過^晶體 °此例示中的切換開關 可以使用其他的替代技央,音耦接至控制信號§1。也 換開關別可 I舉例而言,切 徑及提供互補的問極控制信號南奴電曰曰體具有平行的導通路 關閉切換開_ 570建立一個自 542至正輸入546的負於 ,异风X态540的輸出 運算放大器540的二$開關570關閉時^ ^ 的閘極,苴可以減!L。、的電a被偏壓至負載電晶體 介於節取樣時間區間中,很快地設定 於卽點572a、572b之間的操作電壓。 540 「以由以下的描述加以解釋。運算放大器 匕=增加會導致輸出542的電壓增加,且因 = Ϊ的電麗增加’而導致負載電晶體530汲極的 hi電降體522、524的源極電壓。如此 雷曰辦m ^ 降了電θ曰體522的汲極電壓。因為 以及極與運算放大器540的正輸入546連接,如此 放大器540正輸入546的電壓降低。如此會與原 十旦的正輸入546電壓改變的方向相反。因而此回授是負 容器C2 574是耦接介於負載電晶體530的閘極終 鈿〃、第一導通終端之間。當切換開關570開啟時,此回授路徑 14 201133504 °572a ^ 至源極終端由此保持載電晶體530的閘極 啟可以防止運算放大ί t建將切換開關5 7 〇開 為節點糾顺間内因 ^!Ιΐ! Z:7: u ^ , ,, , — n 不D/Zb之間。此保持電容器C2 574 雷ΪΪ::閘極至源極電壓於負載電晶體530,所以通過 ”電,體>530的電流在保持區間保持大致上是一定值。可以The bias voltage is maintained between the gates of the transistors 522, 524, so the transistors 522, 524 remain on, and the current difference between the inrush current IsiNK and the reference current IREF (if any) The interval remains approximately constant. It is understood that maintaining the bias voltage by the holding capacitor C1 525 between the gate terminal and the source terminal of the transistors 522, 524 may be due to charge in the hold time interval. The leakage and the instantaneous noise of the switch 526 are turned on. The load transistor 530 has a drain terminal coupled to the primary terminal of the transistors 522, 524. The load transistor 530 has a source terminal and a negative rise. The voltage regulator/circuit 532 is coupled 'which provides a negative bias voltage. The current through the load transistor 530 is a combination of the reference electrical Iref of the transistor 522 and the inrush current ISENK of the transistor 524. The current of the transistor = 530 is obtained by the load transistor 530 converting the voltage of one of the terminals of the transistors 522, 524: the sum of the sample and hold j: and the current input circuit 31 can also include an operational amplifier 540. The supply voltage VDD and the negative booster circuit 532 provide a bias voltage to the operational amplifier 540. In this illustration, the operational amplifier 54A has a negative 13 201133504 input 544 that is grounded to ground. This operational pay line 346a is coupled. The effect of the two sides of the positive input 540 and the reference voltage 546 and the reference ground on the reference reference line 346a is the ground potential. This noun, the large f, the ^EF ^ amp 540 bias and 546 Because the operation is placed to represent the difference between the two input terminals 544, the voltage is generated by a non-zero input offset. The load is made by the switch 57 and the 570 is a single pass. The switching switch in the illustration can use other alternative technologies, and the sound is coupled to the control signal § 1. Also, the switch can be used as an example. The cut diameter and the complementary gate signal are provided in parallel. The turn-off path of the turn-off switch _ 570 establishes a negative from 542 to the positive input 546, and the output of the operational wind amplifier 540 of the off-air X state 540 is turned off. The gate of the ^ ^ is turned off, 苴 can be reduced! L. Electric a is biased to the load transistor In the section sampling time interval, the operating voltage between the points 572a, 572b is quickly set. 540 "Explained by the following description. The increase in the operational amplifier 匕 = causes the voltage of the output 542 to increase, and = Ϊ The galvanic voltage is increased, which results in the source voltage of the hi-electric drop bodies 522, 524 of the load transistor 530 bungee. Thus, the thunder voltage of the θ body 522 is reduced by the θ 曰 body 522. The positive input 546 of 540 is connected such that the voltage at amplifier 540 positive input 546 is reduced. This will be the opposite of the original tenant's positive input 546 voltage change. Thus, the feedback is that the negative container C2 574 is coupled between the gate terminal of the load transistor 530 and the first conductive terminal. When the switch 570 is turned on, the feedback path 14 201133504 °572a ^ to the source terminal thereby maintaining the gate of the carrier 530 can prevent the operation amplification, and the switch 5 7 is opened to the node correction room. Internal Cause ^!Ιΐ! Z:7: u ^ , ,, , — n Not between D/Zb. The holding capacitor C2 574 is thundered: the gate-to-source voltage is applied to the load transistor 530, so the current through the "electric, body > 530 remains substantially constant over the hold interval.

^解的疋’郎點572a和572b之間的操作電壓及因此負載電晶 的p 1在保持時間區間或許相為電荷$漏以及切換 開關526開啟的瞬間雜訊而些許改變。 °°° 〇的感測放大器571具有一參考輸入或節點595與 多考線346a耦接,且具有一感測輸入或節點59〇與資料線332& 耦接。如同之前所描述的,參考線346a上的參考電壓及參考 輸入595是由運算放大器54〇偏壓至Vtref,在此範例十是為 地電位。 感測電流iSENSE,其為讀取電流Icell與該汲入電流Isink 之間的一差值,提供至感測放大器571的感測節點59〇。此感 測電流ISENSE係在保持時間區間由將等效負載電容Q(3adi充電 而轉換成感測節點590的一電壓。感測放大器570係響應感測 輸入590與參考輸入595之間的電壓差值,而產生用來指示儲 存於該選取記憶胞510中的一輸出信號576,其係由感測致能 信號SEN所觸發。 放電電晶體580與資料線332a辆接。放電電晶體580的 閘極耦接至一放電2信號582,其用來將資料線332a耦接, 且因此將感測輪入590接地。替代地,也可以使用不是地的偏 壓電壓。 放電電晶體584與參考線346a耦接。放電電晶體584的 閘極耦接至一放電1信號586,其用來將參考線346a耦接, 15 201133504 且因此將參考輸入595接地。替代地,也可以使用不是地的偏 壓電壓。 第6圖顯示操作第5圖中所示的取樣和保持之汲入電流電 路310,以感測儲存在此所選取記憶胞51〇中的資料值之一時 序圖。可以理解的是,第6圖中的時序圖是經過簡化的且並未 等比例繪示。 、在時間T0,放電1信號586被發出以開啟放電電晶體 並與參考線346a耦接,且因此將參考輸入595接地。放電2 信號582被發出以開啟放電電晶體58〇並與資料線332a耦 接,且因此將感測輸入590接地。 在時間T1,列解碼器322響應地址信號而施加一讀取偏 # 壓VwwlEAD至與該被選取之記憶胞510閘極耦接的字元線 324a ’行解碼器326響應地址信號而施加一讀取偏壓Vblread 至與該被選取之記憶胞510汲極終端511耦接的位元線328a, 及將位元線328b與資料線332a耦接。施加至字元線\24a及 位元線328a的偏壓誘發一讀取電流IcELL自汲極終端511至源 極終端512 ’且進入位元線328b與資料線332a。 ' 如第6圖所示’放電2信號582持續被發出所以放電電晶 體580持續開啟,且資料線332a及感測輸入59〇持續在介: 時間T1與T2的取樣區間保持接地。 ’ ;1、 列解碼态344響應地址信號而施加一讀取偏壓腳至聲 與該被選取之參考胞560閘極耦接的字元線345a,器 342響應地址信號而施加一讀取偏壓Vblref至與該被選取^ 參考胞560汲極終端561耦接的位元線343a ’及將位"元線343b 與資料線346a耦接。施加至字元線345a及位元線343a 壓誘發一參考電流IREF自汲極終端561至源極終端562 7至: 元線343b而進入字元線346a。 立 放電1信號586關閉放電電晶體584,且控制信穿 發出以關閉切換開關526和570。如此會建立此取樣和保持子 16 201133504 的操作點,其包括設定介於節點572一 曰體=ϊϊϊί f6可以使電流鏡52G中二極體式連接的電 曰中曰ί =輕接以純參考電流w電流鏡520 晶體522中的參考電流w以導从及入電流1·以響應電 至正可以建立此運算放大器540的輸請 負回授路徑。由運算放大器540的輸出542所 Ϊ = 壓至節點心’以設定介於節點572a、572b 極i端=”係連接在負载電晶體530的閘極終端與源 屯&差且強迫正輸入546的電壓大致與負 目等以減少其鍊。如此允許此取樣和保持之 /及入電机電路310的操作點很快地建立。 获由Ϊ i 2 Ϊ ί連接的電晶體522之外,關閉切換開關526 =广電阳體522的閘極至汲極終端之直接路徑以減少 j杈路㈣封閉迴路反應時間。此外,電容器α仍提供 晶? 522、524閉極終端至第二導通終端之AC信號 、八、,、5路徑,其也減少此回授路徑的封閉迴路反應時間。 換^關,和電容器C1 525及C2 574也提供運算放大 ^路徑的穩定性。此回授路徑的穩定性係與自正輸入 包括三個極點’―個最重要極點在運算放第大器圖5:〇 ϋ及因為電容器C1 525矛口切換開關526的極點。在 542的極點是最重要的是因為運算放大器 RC ^ f53G的閘極電谷及電容器C2…的相對大等效電容 525和切換開關526也提供電晶體522源極至汲極之 #號的正向供給路徑,所以電晶體522汲極至源極的等效 17 201133504 電阻是相對小的。其結果是,電容 造成的極點是遠離最重要極點, 5和切換開關526 性。 具改善了此封閉迴路的穩定 在此情況下,於自時間Ti $ 取樣和保持之沒入電流電路31〇的―2 2^區間。可以達成此 一但使用此回授路徑建立此取 ^^阿速彳呆作點。 310的操作點之後,在時間τ 2此‘ 3持之》及入電流電路 ""526The operating voltage between the 疋 郎 朗 points 572a and 572b and thus the p 1 of the load cell crystal may change slightly during the hold time interval, perhaps as the charge $drain and the instant noise of the switch 526. The sense amplifier 571 has a reference input or node 595 coupled to the multi-test line 346a and has a sense input or node 59 〇 coupled to the data line 332 & As previously described, the reference voltage on reference line 346a and reference input 595 are biased by operational amplifier 54 to Vtref, which in this example is ground. The sense current iSENSE, which is a difference between the read current Icell and the inrush current Isink, is provided to the sense node 59A of the sense amplifier 571. The sense current ISENSE is converted into a voltage of the sense node 590 by charging the equivalent load capacitance Q (3adi) during the hold time interval. The sense amplifier 570 is responsive to the voltage difference between the sense input 590 and the reference input 595. The value is generated to indicate an output signal 576 stored in the selected memory cell 510, which is triggered by the sensing enable signal SEN. The discharge transistor 580 is coupled to the data line 332a. The gate of the discharge transistor 580 The pole is coupled to a discharge 2 signal 582 for coupling the data line 332a and thus grounding the sense wheel 590. Alternatively, a bias voltage other than ground can also be used. Discharge transistor 584 and reference line 346a is coupled. The gate of the discharge transistor 584 is coupled to a discharge 1 signal 586, which is used to couple the reference line 346a, 15 201133504 and thus ground the reference input 595. Alternatively, it is also possible to use a bias other than ground. Fig. 6 shows a timing chart for operating the sample and hold current input circuit 310 shown in Fig. 5 to sense a data value stored in the selected memory cell 51〇. It is understood that , when in Figure 6 The sequence diagram is simplified and not shown to scale. At time T0, a discharge 1 signal 586 is asserted to turn on the discharge transistor and coupled to reference line 346a, and thus ground reference input 595. Discharge 2 signal 582 Is emitted to turn on the discharge transistor 58A and coupled to the data line 332a, and thus ground the sense input 590. At time T1, the column decoder 322 applies a read bias voltage VwwlEAD to the address in response to the address signal. The selected memory cell 510 gate coupled word line 324a 'the row decoder 326 applies a read bias voltage Vblread to the bit line 328a coupled to the selected memory cell 510 drain terminal 511 in response to the address signal. And coupling the bit line 328b to the data line 332a. The bias applied to the word line \24a and the bit line 328a induces a read current IcELL from the drain terminal 511 to the source terminal 512' and enters the bit Line 328b and data line 332a. ' As shown in Fig. 6, 'discharge 2 signal 582 continues to be emitted so that discharge transistor 580 is continuously turned on, and data line 332a and sense input 59 〇 continue to be sampled: time T1 and T2 are sampled The interval remains grounded. ' ;1 column decoding state The 344 applies a read bias pin to the word line 345a coupled to the gate of the selected reference cell 560 in response to the address signal, and the device 342 applies a read bias voltage Vblref to the selected address in response to the address signal. ^ The bit line 343a' coupled to the cell 560 drain terminal 561 and the bit line 343b are coupled to the data line 346a. The word line 345a and the bit line 343a are applied to the voltage source to induce a reference current IREF. The terminal 561 to the source terminal 562 7 to: the source line 343b enter the word line 346a. The discharge 1 signal 586 turns off the discharge transistor 584 and the control signal is issued to turn off the switches 526 and 570. This will establish the operating point of this sample and hold sub 16 201133504, which includes setting the node 572 a body = ϊϊϊί f6 to make the diode in the current mirror 52G connected to the diode 曰 ί = lightly connected to the pure reference current w The reference current w in the current mirror 520 crystal 522 can be used to establish the negative feedback path of the operational amplifier 540 by directing the incoming and outgoing currents 1· in response to the electrical current. The output 542 of the operational amplifier 540 is Ϊ = pressed to the node core 'to set the node 572a, 572b to the end i = " is connected to the gate terminal of the load transistor 530 and the source 屯 & and forced positive input 546 The voltage is approximately equal to the negative, etc. to reduce its chain. This allows the sampling and holding/and operation of the motor circuit 310 to be established quickly. Switching off the transistor 522 connected by Ϊ i 2 Ϊ ί Switch 526 = the direct path of the gate to the drain terminal of the galvanic body 522 to reduce the closed loop reaction time of the 杈 (4). In addition, the capacitor α still provides the AC of the 522, 524 closed-end terminal to the second conductive terminal. The signal, eight, , and five paths also reduce the closed loop response time of this feedback path. Switching, and capacitors C1 525 and C2 574 also provide operational amplification path stability. Stability of this feedback path The system and the self-positive input consist of three poles 'the most important pole in the operation of the amplifier Figure 5: 〇ϋ and because the capacitor C1 525 spear switch 526 pole. The pole at 542 is the most important because of the operation Gate IGBT ^ f53G The relatively large equivalent capacitance 525 of the valley and capacitor C2... and the diverter switch 526 also provide a positive supply path for the source to the drain of the transistor 522, so the equivalent of the transistor 522 to the source of the source 17 201133504 resistance It is relatively small. As a result, the pole caused by the capacitor is far from the most important pole, 5 and the switch 526. It improves the stability of this closed loop. In this case, the sample and keep-off from time Ti $ The current circuit 31〇's “2 2^ interval” can be achieved by using this feedback path to establish this access point. After the operation point of 310, at time τ 2, this '3 hold' and Current circuit ""526

526 ^ J 體522、524閘極終端的偏塵電壓: 二、=住電晶 源極偏麗給電晶體522、524,= fb 上,的閘極至 之後的伴掊卩1匕電日日體522、524在時間T2 ㈣保持Q間保持開啟。可以理解的是, 52 5所保持住的偏壓電壓在保持時間區間或許會^電^ 以及切換開關526開啟的瞬間雜訊而些許改變。' 接,Ϊ = = 出且Τ自控制節點5723解除連 ?, ^ n± . ^ '杈路仫失此,且可以防止控制節點572a電壓 ^生雜訊。介於節點他、之間的操 ^保持電谷裔C2 574所保持。保持電容器 C2 574施力: 土本上固定的閘極至源極偏壓給負載電晶體530,如此負 載電,體530的電流在保持區間保持基本上是固定的。可以理 解的疋,由控制節點572a所保持住的操作電壓,及因此負載 電晶體530的電流,在此保持時間區間或許會因為電荷洩漏以 及切換開關570開啟的瞬間雜訊而些許改變。526 ^ J body 522, 524 gate terminal dust voltage: Second, = live electric crystal source is extremely polarized to the transistor 522, 524, = fb, the gate to the next 掊卩 1 匕 electricity day body 522, 524 keep Q open between times at time T2 (four). It can be understood that the bias voltage held by the 52 5 may change slightly during the hold time interval and the noise of the switch 526 is turned on. ' 接, Ϊ = = 出 Τ 控制 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 572 Between the node and his, the operation is maintained by the electric C2 574. The holding capacitor C2 574 applies a force: the gate-to-source bias fixed to the load cell 530 is so negative that the current of the body 530 remains substantially constant during the hold interval. It can be understood that the operating voltage held by the control node 572a, and thus the current of the load transistor 530, may vary somewhat during this hold time interval due to charge leakage and transient noise of the switch 570 being turned on.

感測電流iSENSE ’其為該讀取電流iCELL與該汲入電流IslNK 之間的一差值,提供至感測放大器571的感測輸入590。 此感測電流ISENSE係由將等效負載電容Ck)adl充電而轉換 成感測輸入590的一電壓。感測輸入590上的電壓會沿著曲線 600假如此被選取記憶胞510是在一較低臨界狀態時,而會沿 著曲線610假如此被選取記憶胞510是在一較高臨界狀態時。 18 201133504 炎本ί時間T3,感測放大器570為了響應與感測輸入590與 二595 ’者之間的電壓差,會產生用來指示該選取記厂虎 L 中的資料值得一輸出信號VOUT。在第6圖中,ν Γ 二第二電壓620假如此被選取記憶胞51〇是在一較低臨:離 第二電壓請假如此被選取記憶胞別是在一“ 因為取樣和保持之汲入電流電路31〇所提供的共同 在於負升壓器532雜訊基本上不會對於感測輸 升懕考輸入595兩者之間的電壓比較有著任何影^。負 56“者雷:生f任何雜訊會導致在記憶胞510與參考胞 I ^改變,所以讀取電流1咖與該沒入電流 夕Γ,κ心二φ ϊ基本上在保持區間是保持基本上不變的。此 朽^持電容㈣仍…⑺施加一個基本上固定的 源極偏壓給電晶體522、524、和530,如此任何注入界 ^的雜訊會導致在閘極電壓類似地改變,所 1源 基本上是保持為定值的,且汲 確性因此’雜訊的存在對比較結果的正 因為感=放大$ 571的感測輸人590之電壓係與讀取電 佈會變得較為緊縮其、。果疋於感測輸入590之電麼分 人59^外而^用曰和保持之沒入電流電路310來偏遷參考輸 工=高^度。當使用參考電流 壓相關聯:因此片,考電麗所需的時間與此參考胞的臨界電 一段時間之後才進行。使用取樣和保 19 201133504 偏壓參考節點,到達參考電麗所需的時^^ I膽來對參考輸入595充電來得快。使用使用參考電流 電路本Λ明第二實關之取樣和保持线入電流 電路310的_要不思圖。在第7圖中’ PMOS電晶體來實施。運算放大器54〇的 * ===與參考…接, 川於輸出542與負輸入544之間的負回授路徑。 第8圖顯示本發明第三實施例之 二 電路31〇的簡要示意圖。在第8圖中,電日f 電流 選擇性地由切換開關526與地輕接 ,二$2 接而不是運算放大器獨的輸人546 也, 耦接的電容來改善此封閉迴路的反應時間。, 、' a 第9圖顯示本發明第四實施例 電路-的簡要示意圖,其包:== === 以將輸人546、544的電壓設定為相it ί?ί 和906由發出信號二: 啟。其:果疋,負輸入544與地耦接而正輸出546與 ^ 耦接/设定通過電容C3 9〇〇的電壓與介於輸入546、54 目同。此切換開關9〇2和9〇6然後被關閉,且開啟切 換開關_。如此將節點9〇8與_接,其提供—個經由’ ^ =3同的專化路,因此設定輸入544的電壓與輸入546的電壓 雖然^發明係已參照實施例來加以描述,然本發明創作並 未受限於其詳細描述内容。替換方式及修改樣式係已於^ 其丨,替換方式及修改樣式將為熟習此項 人士=心及。特別疋’所有具有實質上相同於本發明之 士 合而達成與本發明實質上相同結果者,皆不麟本發明之精g 範t ’所有此料換方式及修改樣式係意欲 於隨附申請專利範圍及其均等物所界定的 本卷月 20 201133504 【圖式簡單說明】 本發明係由申清專利範圍所界定。 特徵,和實施例’會在下列實施方式它目的’ 描述,其中: 早即中搭配圖式被 第1圖顯示使麵極端感測機制的傳統記憶裝置 圖 之示意 係圖 關 第2圖為感測放大器之感測輸入的電壓改變與時間之 第3圖顯示一積體電路的簡要方塊示音 =描述之使用於源極端感測技術的取:和保持㈡ 之一 Ϊ法4^貞圖示此記輯列中的—選取記憶胞進行感測操作 第5圖顯示本發明第一實施例之使用於— 取樣,保持之汲入電流電路的簡要示意圖。。、 ^圖顯示操作第5圖中所示的取樣和保持之沒人電流電 路之時序圖。The sense current iSENSE', which is a difference between the read current iCELL and the sink current IslNK, is provided to the sense input 590 of the sense amplifier 571. This sense current ISENSE is converted to a voltage of the sense input 590 by charging the equivalent load capacitance Ck)adl. The voltage across sense input 590 will be along curve 600 such that memory cell 510 is selected to be in a lower critical state, and along curve 610, if memory cell 510 is selected to be in a higher critical state. 18 201133504 In time T3, in response to the voltage difference between the sense amplifier 590 and the two 595', the sense amplifier 570 is generated to indicate that the data in the selected recorder L is worth an output signal VOUT. In Fig. 6, ν Γ 2 second voltage 620 is thus selected as the memory cell 51 〇 is at a lower side: the second voltage is taken off so that the selected memory cell is in a “because of the intrusion of sampling and retention The commons provided by the current circuit 31〇 is that the negative booster 532 noise does not substantially have any effect on the voltage comparison between the sensed input and sense input 595. Negative 56 “Thunder: Any F The noise will cause the memory cell 510 to change with the reference cell I^, so the reading current 1 and the immersed current Γ, κ 二 φ ϊ are substantially constant in the holding interval. This annihilation capacitor (4) still... (7) applies a substantially fixed source bias voltage to transistors 522, 524, and 530, so that any noise injected into the gate will cause the gate voltage to change similarly. The above is kept constant, and the accuracy is therefore 'the existence of noise. The positive result of the comparison result = the amplification of the voltage of 571. The voltage of the input 590 and the reading of the cloth will become tighter. If the input 590 is sensed, the current input circuit 590 and the held current circuit 310 are used to offset the reference input = high ^ degrees. When using the reference current voltage correlation: therefore, the slice, the time required for the test and the critical power of the reference cell are not performed for a while. Using the Sampling and Protection 19 201133504 bias reference node, the reference input 595 is charged quickly when the time required to reach the reference is obtained. Using the reference current circuit, the second practical off-sampling and holding-in current circuit 310 is not to be considered. It is implemented in the PMOS transistor in Fig. 7. The *=== of the operational amplifier 54〇 is connected to the reference, and the negative feedback path between the output 542 and the negative input 544. Fig. 8 is a schematic view showing the second circuit 31A of the third embodiment of the present invention. In Fig. 8, the electric day f current is selectively connected to the ground by the switch 526, and the second $2 is connected instead of the input of the operational amplifier 546. The coupled capacitor is used to improve the reaction time of the closed loop. , 'a Figure 9 shows a schematic diagram of the circuit of the fourth embodiment of the present invention, which includes: == === to set the voltage of the input 546, 544 to phase ί?ί and 906 to signal two : Kai. It is: the negative input 544 is coupled to the ground and the positive output 546 is coupled to the ^. The voltage across the capacitor C3 9〇〇 is the same as the input 546, 54. The switch 9〇2 and 9〇6 are then turned off and the switch _ is turned on. Thus, node 9〇8 is connected to _, which provides a specialization path through '^=3, so the voltage of input 544 and the voltage of input 546 are set. Although the invention has been described with reference to the embodiments, The inventive creations are not limited by the detailed description thereof. The alternatives and modifications to the styles are already in place, and the alternatives and modifications will be familiar to this person. In particular, all of the materials that are substantially identical to the present invention to achieve substantially the same results as the present invention are not intended to be used in the present invention. This patent is defined by the scope of the patent and its equivalent. 20 201133504 [Simplified description of the drawings] The present invention is defined by the scope of the patent application. The features, and embodiments, will be described in the following embodiments, in which: the early matching pattern is shown in Figure 1 by the schematic diagram of the conventional memory device diagram of the surface extreme sensing mechanism. The voltage change of the sense input of the amp and the time of the third figure shows a brief block of the integrated circuit. The description is used to describe the source-end sensing technique: and hold (b) one of the four methods In the column of the record, the memory cell is selected for the sensing operation. Fig. 5 is a schematic view showing the circuit for sinking and holding the current sample used in the first embodiment of the present invention. . The ^ diagram shows the timing diagram of the sampled and held no-current circuit shown in Figure 5.

第7圖顯示本發明第二實施例之使用於—記憶陣列中的 取樣和保持之汲入電流電路的簡要示意圖。 “ 第8圖顯示本發明第三實施例之使用於—記憶陣列中的 取樣和保持之汲入電流電路的簡要示意圖。 第9圖顯示本發明第四實施例之使用於—記憶陣列中的 取樣和保持之汲入電流電路的簡要示意圖。 【主要元件符號說明】 110、510 :選取記憶胞 112 ' 212、511 :源極終端, 114、214、512 :汲極終端 120、220 :字元線 130、132、230、232 :位元線 21 201133504 140、142 :行選擇電晶體 150 :資料線 160 :參考線 170、544、571 :感測放大器 172、590 :感測輸入 174、595 :參考輸入 176、596 :感測輸出 210 :參考胞 300 :積體電路 310、520 :取樣和保持之汲入電流電路 320 :記憶陣列 324 :字元線 322、344 :列解碼器 328 :位元線 332 :資料匯流排 326、342 :行解碼器 330 :源極端感測放大器/資料輸入結構 340 :參考陣列 343 :參考位元線 345 :參考字元線 346、350 :匯流排 352 :資料輸入線 354 :資料輸出線 360 :其他電路 368 :調整偏壓供應電壓 369 :控制器 522、524 :電晶體 525 :保持電容器C1 526、570 :切換開關 530 :負載電晶體 22 201133504 532 :負升壓器 544 :負輸入 546 :正輸入 560 :參考胞 561 :參考胞之汲極終端 562 :參考胞之源極終端 572a、572b :節點 574 :保持電容器C2 576 :輸出信號 580、584 :放電電晶體 586 :放電1 582 :放電2Fig. 7 is a schematic view showing a circuit for sinking and holding the sampling and holding in the memory array according to the second embodiment of the present invention. Fig. 8 is a schematic view showing a sampling and holding current input circuit used in a memory array according to a third embodiment of the present invention. Fig. 9 is a view showing sampling used in a memory array according to a fourth embodiment of the present invention. A schematic diagram of the current input circuit. [Main component symbol description] 110, 510: Select memory cell 112 '212, 511: source terminal, 114, 214, 512: drain terminal 120, 220: word line 130, 132, 230, 232: bit line 21 201133504 140, 142: row select transistor 150: data line 160: reference line 170, 544, 571: sense amplifier 172, 590: sense input 174, 595: reference Inputs 176, 596: Sensing Output 210: Reference Cell 300: Integrated Circuit 310, 520: Sampling and Holding Inrush Current Circuit 320: Memory Array 324: Word Lines 322, 344: Column Decoder 328: Bit Line 332: data bus 326, 342: row decoder 330: source extreme sense amplifier / data input structure 340: reference array 343: reference bit line 345: reference word line 346, 350: bus 352: data input line 354: data output line 360: other electricity 368: Adjusting the bias supply voltage 369: controller 522, 524: transistor 525: holding capacitor C1 526, 570: switching switch 530: load transistor 22 201133504 532: negative booster 544: negative input 546: positive input 560 : reference cell 561: reference cell's drain terminal 562: reference cell source terminal 572a, 572b: node 574: holding capacitor C2 576: output signal 580, 584: discharge transistor 586: discharge 1 582: discharge 2

Claims (1)

201133504 七、申請專利範圍·· 1. 一種5己憶裝置,包含: 一記憶陣列,具有一資料線及一參考電流線; 一 φ二t入電流電路,包括經由一切換開關與該資料線耦接的 一電谷器,以及 一感測放大電路與該資料線及該參考電流線耦接。 2;如申請專利範圍第1項所述之記憶裝置,其中該汲入電流 電路更,括經,一第二切換開關選擇性地與該參考電流線^ 接的一第二電容器。 3. -·如甲M辱利範圍第1項所述之記憶裝置,更包含一夂老雷 流源安排以提供一參考電流至該參考電流線,且其中〆 該記憶陣列安排以提供—讀取電流自 列、 取記憶胞至該資料線; T 雷可響應一介於第一與第二節點間的操作 括-回授路f係響應該參考電流的大小而可用來】;S5 第Z與第;^卩關_操作電壓,該電容11輪在該第-盘第 === 桑作電壓,及該切換開關可用來在-第-時間£間中致i相授路徑,及在該第一時間 二時間區間中失能該回授路徑;以及 該感,放大電路響應於該第二時間區間 該汲入電流之間的差值,而產座一决指示孩讀取電机,、 憶胞中的㈣值之輸$錢。 …存於該被選取記 4包括如巾請專利範_3項所述之記憶裝置,料紐人電流源 201133504 处二f流鏡’其與該資料線及該參考電流源輪接,該電流鏡 參考電流源f收該參考電流,且能響應於該參 考電"α大小而自έ亥資料線導入該汲入電流; 二^電晶體’具有第—及第二導通終端和一控制路端, 電抓及m電μ,其中该電容器轉接 通終端之間以在該第二時間區間中 與第j 電晶體上;以及 ntp保持•作電壓於該負裁 一運算放大器,該運算放*哭1古 结 _接’-第二輸入與該參考電流第工與= 性地經由該切換開關與該負載電晶體的該控制^端·^出選擇 5.如申請專利範圍第4項所述之記憶裝置, 該電流鏡包含第-電晶體及第二電’每 控制終端及第一和第二導通終端,哕 一具有包含一 與該第二電晶體的該控制終端_1第5 = = H 通終端與該參考電流源輕接,且^體的该第-導 端與該第二電晶體的該第二導通終端麵接曰日體的該第-導通終 δ亥負載電晶體的該第一導通終盥 體的該第二導通終端耦接;以及細/、该弟一及該第二電晶 該汲入電流電路更包含一第二 第二電晶體的該控制終端與該==耦接在該第一及該 導通終端之間。 μ第一電晶體的該第二 6. 如申請專利範圍第4項所述之記憶 路更包含-第二切換_可在 中觀入電流電 -及該第二電晶體的該控制終端轉接^間用來將該第 巧第f時間區間中操作用來將該第一及該電壓’且可在 自該第一偏壓電歷解除執接。 〜電曰曰體的該控制終 25 201133504 7.如申請專利範圍第6項所述之記憶裝置,其中該第一及該第 二電晶體的該控制終端選擇性地經由該第二切換開關與該運算 大器的該第二輪入執接。 8· 一種感測—記憶胞的方法,該方法包含: 施加一偏壓至該記憶胞以自該記憶胞誘發一讀取電流至 一資料線; 自一參考電流源提供一參考電流; . 響應一介於第一與第二節點間的操作電壓的大小而自兮 資料線導入一汲入電流,包括: 使用一回授路徑於一第一時間區間中響應該參考電流的 大小而可用來設定介於該第一與第二節點間的該操作電壓;以 及 在該第一時間區間之後的一第二時間區間中保持介於該 第一與第二節點間與該回授路徑無關的該操作電壓;以及 一根據該讀取電流與該汲入電流之間的一差值決定在該第 一時間區間中儲存於該記憶胞中的一資料值。 9.201133504 VII. Patent Application Range·· 1. A 5 memory device comprising: a memory array having a data line and a reference current line; a φ2t input current circuit including coupling to the data line via a switch Connected to a battery, and a sense amplifier circuit coupled to the data line and the reference current line. 2. The memory device of claim 1, wherein the inrush current circuit further comprises a second switch selectively coupled to the reference current line. 3. The memory device of claim 1, wherein the memory device of the first aspect is further configured to provide a reference current to the reference current line, and wherein the memory array is arranged to provide - read Taking the current from the column and taking the memory cell to the data line; T Ray can be used in response to an operation between the first node and the second node, the feedback path f can be used to respond to the magnitude of the reference current]; S5 Z and The first is the operating voltage, the 11th round of the capacitor is at the first disc === the voltage of the sang, and the switch can be used to cause the i-phased path during the -first time, and in the Disabling the feedback path in a time interval; and the sense, the amplification circuit responds to the difference between the inrush currents in the second time interval, and the production seat instructs the child to read the motor, and recalls The (four) value of the cell loses $ money. ... stored in the selected memory 4 includes a memory device as described in the patent application model _3, the source current source 201133504 at the second f flow mirror 'which is connected to the data line and the reference current source, the current The mirror reference current source f receives the reference current, and can introduce the inrush current from the data line in response to the reference power "α size; the second transistor has a first and second conduction terminal and a control circuit End, the electric catch and the m electric μ, wherein the capacitor is turned on between the terminals to be in the second time interval with the jth transistor; and the ntp is maintained at the negative cut-off operational amplifier, the operation put *Cry 1 ancient knot _ 接 '- the second input and the reference current work and = through the switch and the load transistor of the control end ^ ^ out selection 5. As claimed in the fourth item In the memory device, the current mirror includes a first transistor and a second electrical terminal, each control terminal and the first and second conductive terminals, and the first terminal has a control terminal 1 including the second transistor. = H-terminal is lightly connected to the reference current source, and the first guide of the body The second conductive terminal of the first transistor of the second transistor is coupled to the second conductive terminal of the first conductive terminal of the first conductive terminal of the cathode; and The control circuit of the first transistor and the second transistor further includes a second second transistor coupled to the == between the first terminal and the conductive terminal. The second 6. The memory path of the first transistor, as described in claim 4, further includes - the second switching - can be switched in the middle of the current - and the control terminal of the second transistor is switched The operation between the first and the voltages is used to deactivate the first and second voltages. The memory device of claim 6, wherein the control terminal of the first and second transistors is selectively coupled to the second switch via the second switch The second round of the computing device is engaged. 8. A method of sensing a memory cell, the method comprising: applying a bias voltage to the memory cell to induce a read current from the memory cell to a data line; providing a reference current from a reference current source; Introducing an inrush current from the data line between the first and second nodes, comprising: using a feedback path to respond to the magnitude of the reference current in a first time interval The operating voltage between the first and second nodes; and maintaining the operating voltage between the first and second nodes independent of the feedback path in a second time interval subsequent to the first time interval And determining a data value stored in the memory cell in the first time interval according to a difference between the read current and the inrush current. 9. 如申請專利範圍第8項所述之方法,更包含:For example, the method described in claim 8 of the patent scope further includes: # #在該第二時間區間中根據該讀取電流與該汲入電流之間的該 =設定該感測節點上之—電壓,以及偏壓—參考節點至一參考 擔括I且其中該決定儲存於該記憶胞中的該資料值的步驟包含根 二感測-闕該電壓與該參考節關該參考電壓的—差值來決 又所儲存的該資料值。 1〇. —種記憶裝置,包含: 驭Φ己憶陣列’可自該記憶陣列中的—選取記憶胞提供-讀 取電流至—資料線; 一參考電流源,可提供一參考電流; 26 201133504 一與該資料線耦接的汲入電流電路,該汲入電流電路可自 該資料線導入一汲入電流,該汲入電流電路可用來在一第一 間區間中致能一回授路徑以建立該汲入電流,及在該第一' 區間之後的一第二時間區間中失能該回授路徑及保持哕、、及B 電流,以及 —感測放大電路與該資料線耦接,該感測放大 土二時間區間中該讀取電流與該汲入電流之間的—差g應於該 —用來指示儲存於該被選取記憶胞中的㈣值之輪出信號而產生## In the second time interval, according to the ratio between the read current and the inrush current, the voltage on the sensing node is set, and the bias voltage - reference node to a reference is I and the decision is made therein The step of storing the data value in the memory cell includes a difference between the voltage and the reference voltage of the reference node to determine the stored data value. 1〇. A memory device comprising: 驭Φ recall array can be selected from the memory array to select a memory cell to read current to the data line; a reference current source can provide a reference current; 26 201133504 An inrush current circuit coupled to the data line, the inrush current circuit can introduce an inrush current from the data line, and the inrush current circuit can be used to enable a feedback path in a first interval Establishing the inrush current, and disabling the feedback path and maintaining the current, and the B current in a second time interval after the first 'interval, and - the sense amplifying circuit is coupled to the data line, The difference-g between the read current and the inrush current in the second time interval of the sensed amplification soil is generated by the -four-valued round-out signal stored in the selected memory cell. 2727
TW99108947A 2010-03-25 2010-03-25 Current sink system based on sample and hold for source side sensing TWI442408B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW99108947A TWI442408B (en) 2010-03-25 2010-03-25 Current sink system based on sample and hold for source side sensing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW99108947A TWI442408B (en) 2010-03-25 2010-03-25 Current sink system based on sample and hold for source side sensing

Publications (2)

Publication Number Publication Date
TW201133504A true TW201133504A (en) 2011-10-01
TWI442408B TWI442408B (en) 2014-06-21

Family

ID=46751256

Family Applications (1)

Application Number Title Priority Date Filing Date
TW99108947A TWI442408B (en) 2010-03-25 2010-03-25 Current sink system based on sample and hold for source side sensing

Country Status (1)

Country Link
TW (1) TWI442408B (en)

Also Published As

Publication number Publication date
TWI442408B (en) 2014-06-21

Similar Documents

Publication Publication Date Title
Chang et al. An offset-tolerant fast-random-read current-sampling-based sense amplifier for small-cell-current nonvolatile memory
CN100524506C (en) Memory array with fast bit line precharge
TWI312154B (en) Multiple state sense amplifier for memory architecture
CN101010750B (en) MRAM sense amplifier having a precharge circuit and method for sensing
KR101294573B1 (en) Controlled value reference signal of resistance based memory circuit
TW514921B (en) Semiconductor memory device
US9524766B2 (en) Symmetrical differential sensing method and system for STT MRAM
Vittoz et al. Analog storage of adjustable synaptic weights
US8837210B2 (en) Differential sensing method and system for STT MRAM
TWI333212B (en) Non-volatile memory with variable gate bias for a reference transistor and method thereof
US9472245B2 (en) Sense amplifier and related method
US20140063923A1 (en) Mismatch Error Reduction Method and System for STT MRAM
Na et al. An offset-canceling triple-stage sensing circuit for deep submicrometer STT-RAM
JP6137522B2 (en) Photoelectric conversion device reset method, photoelectric conversion device, photoelectric conversion array, and imaging device
TW201113899A (en) Current sink system for source side sensing
US20220157351A1 (en) Sense amplifier and operating method for non-volatile memory with reduced need on adjusting offset to compensate the mismatch
TW200845005A (en) Sensing device for floating body cell memory and method thereof
WO2014158414A1 (en) Low-power, high-accuracy current reference for highly distributed current references for cross point memory
EP2912662A1 (en) Low voltage current reference generator for a sensing amplifier
US9165613B2 (en) Sample-and-hold current sense amplifier and related method
CN105185404A (en) Charge transfer type sense amplifier
Sakimura et al. A 512kb cross-point cell MRAM
TW201133504A (en) Current sink system based on sample and hold for source side sensing
CN110648706B (en) Three-dimensional resistive random access memory and reading circuit thereof
CN102208207A (en) System for sensing sink current based on sampled and maintained source cathode terminals