TWI788987B - Memory cell for analog content-addressable memory and analog content-addressable memory device - Google Patents
Memory cell for analog content-addressable memory and analog content-addressable memory device Download PDFInfo
- Publication number
- TWI788987B TWI788987B TW110132411A TW110132411A TWI788987B TW I788987 B TWI788987 B TW I788987B TW 110132411 A TW110132411 A TW 110132411A TW 110132411 A TW110132411 A TW 110132411A TW I788987 B TWI788987 B TW I788987B
- Authority
- TW
- Taiwan
- Prior art keywords
- type transistor
- channel nand
- sense amplifier
- gate
- nand series
- Prior art date
Links
Images
Landscapes
- Analogue/Digital Conversion (AREA)
- Circuits Of Receivers In General (AREA)
Abstract
Description
本發明是有關於一種用於內容可定址記憶體的記憶胞以及內容可定址記憶體裝置,且特別是有關於一種用於類比內容可定址記憶體的記憶胞以及類比內容可定址記憶體裝置。The present invention relates to a memory cell for a content-addressable memory and a content-addressable memory device, and more particularly to a memory cell for an analog content-addressable memory and an analog content-addressable memory device.
隨著記憶體技術的發展,一種內容可定址記憶體(Content-Addressable Memory, CAM)係被提出。CAM是一種應用於高速記憶體內部(In-memory)搜尋的特殊記憶體,並可以使用高度並行的方式將輸入搜索詞與陣列中所有列的儲存詞進行比較。CAM在圖像(pattern)匹配與搜索的許多應用提供了非常強大的功能。With the development of memory technology, a Content-Addressable Memory (CAM) system is proposed. CAM is a special memory used for in-memory search, and can compare the input search word with the stored words in all columns in the array in a highly parallel manner. CAM provides very powerful functions in many applications of image (pattern) matching and searching.
相較於傳統的三元CAM(ternary CAM),類比CAM顯著地增加了資料密度,並減少記憶體內部之處理電路操作之能量損耗與處理電路之面積。類比CAM 需要具有良好記憶胞的穩定性與較高的陣列密度。隨著大數據(big data)的發展,在龐大資料庫中進行資料搜尋與資料比對時,則需要一個高密度的類比CAM。如何在判斷搜尋範圍是否匹配於儲存範圍時,避免資料比較的誤判,乃業界所致力的方向之一。Compared with the traditional ternary CAM (ternary CAM), analog CAM significantly increases the data density, and reduces the energy consumption and the area of the processing circuit in the memory. Analog CAM requires good memory cell stability and high array density. With the development of big data, a high-density analog CAM is required when searching and comparing data in a huge database. How to avoid misjudgment in data comparison when judging whether the search range matches the storage range is one of the directions that the industry is working on.
本發明係有關於一種用於類比內容可定址記憶體的記憶胞以及類比內容可定址記憶體裝置,其利用電流控制電路來固定匹配範圍內之不同輸入訊號所對應的電流位準,減少資料比較時的誤判情況。The present invention relates to a memory cell used for an analog content addressable memory and an analog content addressable memory device, which uses a current control circuit to fix the current levels corresponding to different input signals within the matching range, reducing data comparison misjudgment situation.
根據本發明之第一方面,提出一種用於一類比內容可定址記憶體(Analog Content-Addressable Memory, analog CAM)的記憶胞。記憶胞包括一N型電晶體、一P型電晶體、及一電流控制電路。N型電晶體具有一第一閘極。N型電晶體的第一閘極用以接收一第一輸入訊號。P型電晶體具有一第二閘極。P型電晶體的第二閘極用以接收一第二輸入訊號。電流控制電路係耦接至N型電晶體及P型電晶體之至少二者之一,用以產生一導通電流。其中,當第一輸入訊號的輸入電壓與第二輸入訊號的輸入電壓位於一匹配範圍內時,N型電晶體與P型電晶體均導通,且對應之導通電流實質上係為一固定電流值。匹配範圍係與N型電晶體的臨界電壓、P型電晶體的臨界電壓、及固定電流值相關。According to a first aspect of the present invention, a memory cell for an Analog Content-Addressable Memory (analog CAM) is proposed. The memory cell includes an N-type transistor, a P-type transistor, and a current control circuit. The N-type transistor has a first gate. The first gate of the N-type transistor is used for receiving a first input signal. The P-type transistor has a second gate. The second gate of the P-type transistor is used for receiving a second input signal. The current control circuit is coupled to at least one of the N-type transistor and the P-type transistor for generating a conduction current. Wherein, when the input voltage of the first input signal and the input voltage of the second input signal are within a matching range, both the N-type transistor and the P-type transistor are turned on, and the corresponding conduction current is substantially a fixed current value . The matching range is related to the critical voltage of the N-type transistor, the critical voltage of the P-type transistor, and the fixed current value.
根據本發明之另一方面,提出一種類比內容可定址記憶體裝置,包括一字元線驅動電路、多個記憶胞、多個匹配訊號線、多個源極線、一源極線驅動電路、及一感測放大器電路。字元線驅動電路用以提供多個第一輸入訊號與多個第二輸入訊號。各記憶胞包括一N型電晶體、一P型電晶體、及一電流控制電路。N型電晶體具有一第一閘極。N型電晶體的第一閘極用以接收對應之第一輸入訊號。P型電晶體具有一第二閘極。P型電晶體的第二閘極用以接收對應之第二輸入訊號。電流控制電路係耦接至N型電晶體及P型電晶體之至少二者之一,用以產生一導通電流。各匹配訊號線係耦接至對應之記憶胞。各源極線係與對應之電流控制電路耦接。源極線驅動電路係耦接至此些源極線。感測放大器電路係耦接至此些匹配訊號線。其中,針對此些記憶胞中之一特定記憶胞,當對應至特定記憶胞之第一輸入訊號的輸入電壓與對應至特定記憶胞之第二輸入訊號的輸入電壓皆位於特定記憶胞之一匹配範圍內時,特定記憶胞之N型電晶體與P型電晶體均導通,且特定記憶胞之導通電流實質上係為一固定電流值。特定記憶胞之匹配範圍係與特定記憶胞之N型電晶體的臨界電壓、特定記憶胞之P型電晶體的臨界電壓、及固定電流值相關。According to another aspect of the present invention, an analog content addressable memory device is proposed, comprising a word line driver circuit, a plurality of memory cells, a plurality of matching signal lines, a plurality of source lines, and a source line driver circuit , and a sense amplifier circuit. The word line driving circuit is used for providing a plurality of first input signals and a plurality of second input signals. Each memory cell includes an N-type transistor, a P-type transistor, and a current control circuit. The N-type transistor has a first gate. The first gate of the N-type transistor is used for receiving the corresponding first input signal. The P-type transistor has a second gate. The second gate of the P-type transistor is used for receiving the corresponding second input signal. The current control circuit is coupled to at least one of the N-type transistor and the P-type transistor for generating a conduction current. Each matching signal line is coupled to the corresponding memory cell. Each source line is coupled with the corresponding current control circuit. The source line driving circuit is coupled to the source lines. The sense amplifier circuit is coupled to the matching signal lines. Wherein, for a specific memory cell among the memory cells, when the input voltage corresponding to the first input signal of the specific memory cell and the input voltage corresponding to the second input signal of the specific memory cell are located in one of the specific memory cells When within the range, both the N-type transistor and the P-type transistor of the specific memory cell are turned on, and the conduction current of the specific memory cell is essentially a fixed current value. The matching range of a specific memory cell is related to the threshold voltage of the N-type transistor of the specific memory cell, the threshold voltage of the P-type transistor of the specific memory cell, and the fixed current value.
根據本發明之再一方面,提出一種類比內容可定址記憶體裝置,包括一第一字元線驅動電路、一第二字元線驅動電路、一第一N通道NAND串列組、一第一P通道NAND串列組、多個第一感測放大電路、多個第二感測放大電路、及多個第一及邏輯閘。第一字元線驅動電路用以提供多個第一輸入訊號,第二字元線驅動電路用以提供多個第二輸入訊號。第一N通道NAND串列組包括多個第一N通道NAND串列。各第一N通道NAND串列用以接收此些第一輸入訊號。各第一N通道NAND串列更用以產生一第一電流。第一P通道NAND串列組包括多個第一P通道NAND串列。各第一P通道NAND串列用以接收此些第二輸入訊號,各第一P通道NAND串列更用以產生一第二電流。此些第一感測放大電路係分別耦接至第一N通道NAND串列組之此些第一N通道NAND串列。此些第二感測放大電路係分別耦接至第一P通道NAND串列組之此些第一P通道NAND串列。此些第一感測放大電路與此些第二感測放大電路係各具有一臨界電流值。各第一及邏輯閘係耦接至對應之第一感測放大電路與對應之第二感測放大電路。其中此些第一及邏輯閘之一為一選定之第一及邏輯閘。當選定之第一及邏輯閘所對應之第一電流與第二電流的電流均大於或等於臨界電流值時,選定之第一及邏輯閘輸出一第一邏輯值。According to still another aspect of the present invention, an analog content addressable memory device is proposed, including a first word line driver circuit, a second word line driver circuit, a first N-channel NAND series group, a first A P-channel NAND series group, a plurality of first sense amplifier circuits, a plurality of second sense amplifier circuits, and a plurality of first and logic gates. The first word line driving circuit is used for providing a plurality of first input signals, and the second word line driving circuit is used for providing a plurality of second input signals. The first N-channel NAND series group includes a plurality of first N-channel NAND series. Each first N-channel NAND series is used for receiving the first input signals. Each first N-channel NAND series is further used to generate a first current. The first P-channel NAND series group includes a plurality of first P-channel NAND series. Each first P-channel NAND series is used to receive the second input signals, and each first P-channel NAND series is further used to generate a second current. The first sense amplifier circuits are respectively coupled to the first N-channel NAND series of the first N-channel NAND series group. The second sense amplifier circuits are respectively coupled to the first P-channel NAND series of the first P-channel NAND series group. Each of the first sense amplifier circuits and the second sense amplifier circuits has a critical current value. Each of the first and logic gates is coupled to a corresponding first sense amplifier circuit and a corresponding second sense amplifier circuit. One of the first and logic gates is a selected first and logic gate. When the currents of the first current and the second current corresponding to the selected first and logic gates are greater than or equal to the critical current value, the selected first and logic gates output a first logic value.
為了對本發明之上述及其他方面有更佳的瞭解,下文特舉實施例,並配合所附圖式詳細說明如下:In order to have a better understanding of the above-mentioned and other aspects of the present invention, the following specific examples are given in detail with the accompanying drawings as follows:
請參照第1圖,其繪示類比內容可定址記憶體(analog content-addressable memory, analog CAM)100之操作示意圖。類比內容可定址記憶體100包括數個類比CAM記憶胞CL1。類比CAM記憶胞CL1排列於多列R(1)、R(2)、R(3)、R(4),以儲存數筆類比內容。舉例來說,儲存於列R(1)的內容是「0.00~1.00、0.48~0.76、0.00~0.15」。「0.00~1.00」代表任何數值都能匹配。儲存於列R(2)的內容是「0.62~1.00、0.25~0.63、0.25~1.00」。儲存於列R(3)的內容是「0.26~0.61、0.12~0.40、0.00~1.00」。儲存於列R(4)的內容是「0.00~0.43、0.00~0.28、0.58~1.00」。數個輸入訊號S1輸入至類比內容可定址記憶體100。第一個輸入訊號S1的內容是「0.81」,第二個輸入訊號S1的內容是「0.62」,第三個輸入訊號S1的內容是「0.12」。Please refer to FIG. 1 , which shows an operation diagram of an analog content-addressable memory (analog CAM) 100 . The analog content
這些輸入訊號S1與儲存於列R(1)的內容進行比對。儲存於列R(1)之「0.00~1.00、0.48~0.76、0.00~0.15」係為匹配。由於「0.81」、「0.62」及「0.12」分別落入「0.00~1.00」、「0.48~0.76」及「0.00~0.15」,故據以輸出匹配成功結果Ry。These input signals S1 are compared with the contents stored in row R(1). "0.00~1.00, 0.48~0.76, 0.00~0.15" stored in column R(1) is a match. Since "0.81", "0.62" and "0.12" fall into "0.00-1.00", "0.48-0.76" and "0.00-0.15" respectively, the successful matching result Ry is output accordingly.
這些輸入訊號S1與儲存於列R(2)之內容進行比對後,輸出匹配不成功結果Rn。這些輸入訊號S1與儲存於列R(3)之內容進行比對後,輸出匹配不成功結果Rn。這些輸入訊號S1與儲存於列R(4)之內容進行比對後,輸出匹配不成功結果Rn。也就是說,類比內容可定址記憶體100可以儲存類比內容,任何與輸入訊號S1匹配的類比內容可以被搜尋出來。After these input signals S1 are compared with the contents stored in the row R(2), a result of unsuccessful matching Rn is output. After these input signals S1 are compared with the contents stored in the column R(3), a result of unsuccessful matching Rn is output. After these input signals S1 are compared with the contents stored in the column R(4), the result of unsuccessful matching Rn is output. That is to say, the analog content
請參照第2圖,其繪示類比CAM記憶胞CL1。類比CAM記憶胞CL1包括一第一浮動閘極裝置MSn及第二浮動閘極裝置MSp。第一浮動閘極裝置MSn具有一N型通道,第二浮動閘極裝置MSp具有一P型通道。第二浮動閘極裝置MSp以串聯之方式連接於第一浮動閘極裝置MSn。第一浮動閘極裝置MSn係為一N型金氧半導體(NMOS),第二浮動閘極裝置MSp係為一P型金氧半導體(PMOS)。第一浮動閘極裝置MSn之汲極連接於一匹配訊號線ML,第一浮動閘極裝置MSn之源極連接於第二浮動閘極裝置MSp之源極。第二浮動閘極裝置MSp之汲極連接於一源極線SL。輸入訊號S1同時輸入至第一浮動閘極裝置MSn與第二浮動閘極裝置MSp之閘極。Please refer to FIG. 2, which shows the analog CAM memory cell CL1. The analog CAM memory cell CL1 includes a first floating gate device MSn and a second floating gate device MSp. The first floating gate device MSn has an N-type channel, and the second floating gate device MSp has a P-type channel. The second floating gate device MSp is connected in series to the first floating gate device MSn. The first floating gate device MSn is an N-type metal oxide semiconductor (NMOS), and the second floating gate device MSp is a P-type metal oxide semiconductor (PMOS). The drain of the first floating gate device MSn is connected to a matching signal line ML, and the source of the first floating gate device MSn is connected to the source of the second floating gate device MSp. The drain of the second floating gate device MSp is connected to a source line SL. The input signal S1 is simultaneously input to the gates of the first floating gate device MSn and the second floating gate device MSp.
請參照第3圖,其繪示類比CAM記憶胞CL1的匹配範圍MR。曲線CN係為第一浮動閘極裝置MSn之特性曲線,曲線CP係為第二浮動閘極裝置MSp之特性曲線。曲線CN與曲線CP之陡峭斜率大於0.01 mV/dec。舉例來說,第3圖之曲線CN及曲線CP的陡峭曲線係為0.015 mV/dec。第一浮動閘極裝置MSn與第二浮動閘極裝置MSp係為超陡峭元件(super steep slope devices)。Please refer to FIG. 3, which shows the matching range MR of the analog CAM memory cell CL1. Curve CN is the characteristic curve of the first floating gate device MSn, and curve CP is the characteristic curve of the second floating gate device MSp. The steep slope of curve CN and curve CP is greater than 0.01 mV/dec. For example, the steep curves of curve CN and curve CP in Fig. 3 are 0.015 mV/dec. The first floating gate device MSn and the second floating gate device MSp are super steep slope devices.
在類比CAM記憶胞CL1中,第一浮動閘極裝置MSn之臨界電壓低於第二浮動閘極裝置MSp之臨界電壓,以於第一浮動閘極裝置MSn之臨界電壓與第二浮動閘極裝置MSp之臨界電壓之間形成匹配範圍MR。在類比CAM記憶胞CL1中,匹配範圍MR之下限LB係為第一浮動閘極裝置MSn之臨界電壓,匹配範圍MR之上限UB係為第二浮動閘極裝置MSp之臨界電壓。In the analog CAM memory cell CL1, the threshold voltage of the first floating gate device MSn is lower than the threshold voltage of the second floating gate device MSp, so that the threshold voltage of the first floating gate device MSn and the second floating gate device The threshold voltages of MSp form a matching range MR. In the analog CAM memory cell CL1, the lower limit LB of the matching range MR is the critical voltage of the first floating gate device MSn, and the upper limit UB of the matching range MR is the critical voltage of the second floating gate device MSp.
如第2圖及第3圖所示,當輸入訊號S1落於匹配範圍MR之內時,第一浮動閘極裝置MSn被導通且第二浮動閘極裝置MSp也被導通,故形成了導通電流Ip。當輸入訊號S1不位於匹配範圍MR之內時,第一浮動閘極裝置MSn不導通或者第二浮動閘極裝置MSp不導通,故不會形成導通電流Ip。As shown in Figure 2 and Figure 3, when the input signal S1 falls within the matching range MR, the first floating gate device MSn is turned on and the second floating gate device MSp is also turned on, thus forming a conduction current Ip. When the input signal S1 is not within the matching range MR, the first floating gate device MSn is not turned on or the second floating gate device MSp is not turned on, so the conduction current Ip will not be formed.
請參照第4圖,其繪示根據本揭露一實施例之類比內容可定址記憶體200的記憶胞202之電路圖。類比內容可定址記憶體200例如包括記憶胞202與匹配訊號線204。記憶胞202包括N型電晶體206、P型電晶體208與電流控制電路210。N型電晶體206具有閘極G1,N型電晶體206的閘極G1用以接收輸入訊號B(i)。P型電晶體208具有閘極G2,P型電晶體208的閘極G2用以接收輸入訊號A(i)。電流控制電路210耦接至N型電晶體206及P型電晶體208之至少二者之一,例如是耦接至P型電晶體208,電流控制電路210並用以產生至少一導通電流,例如是導通電流I
pass。其中,當輸入訊號B(i)的輸入電壓與輸入訊號A(i)的輸入電壓位於匹配範圍內時,N型電晶體206與P型電晶體208均導通,且對應之導通電流I
pass實質上係為一固定電流值。匹配範圍係與N型電晶體的臨界電壓206、P型電晶體208的臨界電壓、及此固定電流值相關。
Please refer to FIG. 4 , which shows a circuit diagram of a
如此,藉由使用電流控制電路210,使得N型電晶體206與P型電晶體208導通時流過N型電晶體206與P型電晶體208的電流大小,被固定為電流控制電路210的電流大小。如此,即使輸入訊號A(i)或B(i)的輸入電壓可能有多個不同的位準,仍使得N型電晶體206與P型電晶體208導通時流過N型電晶體206與P型電晶體208的電流大小實質上相同。如此,即使輸入訊號A(i)或B(i)的輸入電壓可能有多個不同的位準,仍可使得連接於記憶胞202之匹配訊號線204上的放電時間相同,以致使匹配訊號線204具有穩定的放電時間以減少誤判產生。In this way, by using the
如第4圖所示,電流控制電路210例如具有控制電晶體210A,控制電晶體210A為金屬氧化物半導體場效電晶體(MOSFET)或浮動閘極(Float gate, FG)電晶體。N型電晶體206之汲極D1係用以與匹配訊號線204電性連接,N型電晶體206之源極S1係與P型電晶體208之源極S2電性連接,P型電晶體208之汲極D2係與控制電晶體210A之一端(例如是控制電晶體210A之汲極D3)電性連接。控制電晶體210A係具有閘極G3,控制電晶體210A之閘極G3係用以接收控制電壓C(i),控制電壓C(i)係實質上為一固定電壓值。As shown in FIG. 4 , the
控制電晶體210A之另一端(例如是控制電晶體210A之源極S3)例如是與一源極線SL電性連接。在一實施例中,N型電晶體206與P型電晶體208係為2D(二維)快閃記憶體結構或3D(二維)快閃記憶體結構。在一實施例中,N型電晶體206與P型電晶體208可以使用F-N穿隧(Fowler-Nordheim tunneling, FN tunneling)、通道熱電子注入技術(Channel Hot Electron, CHE)或是多晶矽對多晶矽(poly to poly)的方式來進行編程(program)。於進行編程時,輸入訊號A(i)的輸入電壓與輸入訊號B(i)的輸入電壓可為不同。而於進行資料搜尋與比較時,輸入訊號A(i)的輸入電壓與輸入訊號B(i)的輸入電壓則為相同。在一實施例中,N型電晶體206與P型電晶體208可以使用F-N穿隧、帶對帶熱電洞入射(Band-To-Band Hot Hole, BTBHH)或是多晶矽對多晶矽的方式來進行抹除(erase)。The other end of the
在一實施例中,類比內容可定址記憶體200係為快閃記憶體。快閃記憶體例如為電荷儲存記憶體(Charge storage memory)、電荷捕捉記憶體(Charge trapping memory)、分離式閘極記憶體(Split gate memory)或鐵電體場效電晶體(Ferroelectric field-effect transistor, FeFET)記憶體。在另一實施例中,類比內容可定址記憶體200係為超陡峭(super steep slope)快閃記憶體。超陡峭快閃記憶體為閘流體隨機存取記憶體(Thyristor Random Access Memory, TRAM)、閘極控制閘流體(Gate Control Thyristor, GCT)、穿隧式場效電晶體(Tunnel Field-Effect Transistor, TFET)或負電容場效電晶體(Negative Capacitance Field-Effect Transistor, NCFET)。In one embodiment, the analog content
請參照第5A圖與第5B圖,第5A圖繪示未具有電流控制電路210之記憶胞202的匹配範圍MR0,第5B圖繪示連接於第5A圖的記憶胞202之匹配訊號線204的輸出電壓與放電時間的關係圖。曲線CN係為N型電晶體206之特性曲線,曲線CP係為P型電晶體208之特性曲線,N型電晶體206的臨界電壓Vthn (例如,約-2.8V)與P型電晶體208的臨界電壓Vthp(例如,約-0.4V)定義出匹配範圍MR0。當N型電晶體206與P型電晶體208操作於次臨界區(subthreshold region)時,即可能產生之曲線CN與CP在匹配範圍MR0中的斜率不夠陡峭的情形。如第5A圖與第5B圖所示,當記憶胞202未具有電流控制電路210時,當輸入訊號A(i)及B(i)的輸入電壓為V1時,N型電晶體206的導通電流為I1,使得P型電晶體208的導通電流亦為I1。此時,導通的N型電晶體206與P型電晶體208將使得匹配訊號線204上的輸出電壓由原始電壓Vms依照曲線V(I1)下降,而於時間點t2下降至用以判斷是否匹配的參考電壓Vref。同理,當輸入訊號A(i)及B(i)的輸入電壓為V2時,N型電晶體206的導通電流為I2,使得匹配訊號線204上的輸出電壓由原始電壓Vms依照曲線V(I2)下降,而於時間點t1下降至用以判斷是否匹配的參考電壓Vref。當輸入訊號A(i)及B(i)的輸入電壓為V3時,N型電晶體206的導通電流為I3,使得匹配訊號線204上的輸出電壓由原始電壓Vms依照曲線V(I3)下降,而於時間點t3下降至用以判斷是否匹配的參考電壓Vref。由於不同的輸入訊號A(i)及B(i)的輸入電壓將對應至不同的導通電流大小,而使得匹配訊號線204上的輸出電壓由原始電壓Vms下降至參考電壓Vref的時間不同。也就是說,對於電壓位準不同的多個輸入訊號A(i)或B(i)來說,由於匹配訊號線204放電時的電流位準並未固定,使得連接於記憶胞202之匹配訊號線204上的放電時間不同,以致使匹配訊號線204具有不同的放電時間。如此,將使得要在進行記憶體內部之資料搜尋與比對,以判斷記憶胞所儲存的資料範圍以及所輸入的資料是否匹配時,進行判斷動作所需的時間長度變得不固定,而使得誤判的機率上升。Please refer to FIG. 5A and FIG. 5B. FIG. 5A shows the matching range MR0 of the
請參照第6A圖與第6B圖,第6A圖繪示具有電流控制電路210之記憶胞202的匹配範圍MR1,第6B圖繪示連接於第6A圖的記憶胞202之匹配訊號線204的輸出電壓與放電時間的關係圖。第6A圖的曲線CN與曲線CP相同於第5A圖的曲線CN與曲線CP。具有電流控制電路210之記憶胞202的匹配範圍MR1係與N型電晶體206的臨界電壓、P型電晶體208的臨界電壓相關、及固定電流值相關。匹配範圍MR1例如是由N型電晶體206與P型電晶體208以導通電流I
pass均導通時,所對應之N型電晶體206的最小的閘極電壓V4(大於N型電晶體206的臨界電壓Vthn),以及P型電晶體208的最大閘極電壓V5(小於P型電晶體208的臨界電壓Vthp)所定義。如第6A圖與第6B圖所示,當輸入訊號A(i)及B(i)的輸入電壓為V1時,N型電晶體206與 P型電晶體208的導通電流為Ipass。此時,導通的N型電晶體206與P型電晶體208將使得匹配訊號線204上的輸出電壓由原始電壓Vms依照曲線V下降,而於時間點t4下降至用以判斷是否匹配的參考電壓Vref。同理,當輸入訊號A(i)及B(i)的輸入電壓為V2及V3時,N型電晶體206的導通電流依然為Ipass,使得匹配訊號線204上的輸出電壓由原始電壓Vms仍然依照曲線V)下降,而分別於時間點t4下降至用以判斷是否匹配的參考電壓Vref。由於不同的輸入訊號A(i)及B(i)的輸入電壓將對應至相同的導通電流Ipass,而使得匹配訊號線204上的輸出電壓由原始電壓Vms下降至參考電壓Vref的放電時間幾乎相同。也就是說,當記憶胞202具有電流控制電路210時,由於匹配訊號線204放電時的電流大小被固定,使得連接於記憶胞202之匹配訊號線204上的不同之輸入訊號的輸入電壓所對應的放電時間實質上相同,以使匹配訊號線204具有穩定的放電時間。如此,將使得要在進行記憶體內部之資料搜尋與比對,以判斷記憶胞所儲存的資料與所輸入的資料是否匹配時,進行判斷動作所需的時間長度係為固定,而減少誤判的機率。
Please refer to FIG. 6A and FIG. 6B. FIG. 6A shows the matching range MR1 of the
請參照第7圖,其繪示將第4圖所示之記憶胞應用於類比內容可定址記憶體裝置300之一例。類比內容可定址記憶體裝置300包括字元線驅動電路302、多個記憶胞304(即第4圖的記憶胞202)、多個匹配訊號線312、多個源極線314、源極線驅動電路316與感測放大器電路(Sense Amplifier)318。字元線驅動電路302用以提供多個輸入訊號B(1), B(2), …, B(n)與多個輸入訊號A(1), A(2, …, A(n)。類比內容可定址記憶體裝置300的各個記憶胞304包括N型電晶體306、P型電晶體308與電流控制電路310,各個記憶胞304的N型電晶體306具有閘極G1,各個記憶胞304的N型電晶體306的閘極G1用以接收對應之輸入訊號B(i),i為1至n之正整數。各個記憶胞304的P型電晶體308具有閘極G2,各個記憶胞304的P型電晶體308的閘極G2用以接收對應之輸入訊號A(i)。電流控制電路310耦接至P型電晶體308,用以產生導通電流I
pass。各個匹配訊號線312係耦接至對應之記憶胞304,各個源極線314係與對應之電流控制電路310耦接。源極線驅動電路316耦接至多個源極線314,感測放大器電路318耦接至多個匹配訊號線312。其中,針對該些記憶胞中之一特定記憶胞304(j, i) (j為介於1與m之間的整數),當對應至特定記憶胞304(j, i)之輸入訊號B(i)的輸入電壓與對應至特定記憶胞304(j, i)之輸入訊號A(i)的輸入電壓皆位於特定記憶胞304(j, i)之匹配範圍內時,特定記憶胞304(j, i)之N型電晶體306與P型電晶體308均導通,且特定記憶胞304(j, i)之導通電流I
pass實質上係為一固定電流值。特定記憶胞304(j, i)之匹配範圍係與特定記憶胞304(j, i)之N型電晶體306的臨界電壓、特定記憶胞304(j, i)之P型電晶體308的臨界電壓、與此固定電流值相關。
Please refer to FIG. 7 , which shows an example of applying the memory cell shown in FIG. 4 to an analog content
當對多個記憶胞304(例如是記憶胞304(1, 1)至記憶胞304(m, n))進行編程或抹除時,輸入訊號A(i)可與輸入訊號B(i)不同。而當使輸入訊號A(1)至A(n)與儲存於第1列至第m列記憶胞304的內容進行比對時,則輸入訊號A(i)係輸入訊號B(i)相同。When programming or erasing a plurality of memory cells 304 (eg memory cell 304(1, 1) to memory cell 304(m, n)), the input signal A(i) may be different from the input signal B(i) . When the input signals A(1) to A(n) are compared with the contents stored in the
舉例來說,當輸入訊號A(1)至輸入訊號A(n)與儲存於第1列至第m列之記憶胞304(1, 1)至記憶胞304(m, n)的內容進行比對時,假設輸入訊號A(1)至輸入訊號A(n)分別位於第1列之記憶胞304(1, 1)至記憶胞304(1, n)之匹配範圍內時,代表輸入訊號A(1)至輸入訊號A(n)的內容與儲存於第1列之記憶胞304(1, 1)至記憶胞304(1, n)的內容為匹配。亦即是輸入訊號A(1)至輸入訊號A(n)所對應之類比數值,係分別位於第1列之記憶胞304(1, 1)至記憶胞304(1, n)之匹配範圍所對應之類比數值範圍之內。此時,第1列之記憶胞304(1, 1)至記憶胞304(1, n)均會導通而有電流流過,以將匹配訊號線312(1)的電壓下拉,使感測放大器電路318偵測出匹配成功結果。For example, when the input signal A(1) to the input signal A(n) are compared with the contents stored in the memory cell 304(1, 1) to the memory cell 304(m, n) in the first column to the mth column Timing, assuming that the input signal A(1) to the input signal A(n) are respectively within the matching range of the memory cell 304(1, 1) to the memory cell 304(1, n) in the first row, it means the input signal A (1) The content of the input signal A(n) matches the content stored in the memory cells 304(1, 1) to 304(1, n) of the first column. That is to say, the analog values corresponding to the input signal A(1) to the input signal A(n) are determined by the matching ranges of the memory cell 304(1, 1) to the memory cell 304(1, n) respectively located in the first column within the corresponding analog value range. At this moment, the memory cell 304(1, 1) to the memory cell 304(1, n) in the first column are all turned on and current flows, so as to pull down the voltage of the matching signal line 312(1), so that the sense amplifier The
而如果輸入訊號A(1)至輸入訊號A(n)有任何一個不位於第1列之記憶胞304(1, 1)至記憶胞304(1, n)之匹配範圍內時,代表輸入訊號A(1)至輸入訊號A(n)的內容與儲存於第1列之記憶胞304(1, 1)至記憶胞304(1, n)的內容為不匹配。亦即是輸入訊號A(1)至輸入訊號A(n)所對應之類比數值,並沒有完全位於第1列之記憶胞304(1, 1)至記憶胞304(1, n)之匹配範圍所對應之類比數值範圍之內。此時,第1列之記憶胞304(1, 1)至記憶胞304(1, n)至少一者不會導通,而使得匹配訊號線312(1)的電壓不會被下拉至小於參考電壓Vref。如此,感測放大器電路318將偵測出匹配不成功結果。也就是說,類比內容可定址記憶體300可以儲存類比內容,任何與輸入訊號A(1)至輸入訊號A(n)匹配的類比內容(例如是某一列之記憶胞304所儲存的類比內容)可以被搜尋出來。And if any one of the input signal A(1) to the input signal A(n) is not within the matching range of the memory cell 304(1, 1) to the memory cell 304(1, n) in the first column, it means the input signal The contents of A(1) to input signal A(n) do not match the contents of memory cells 304(1, 1) to 304(1, n) stored in
藉由使用第4圖所示之記憶胞,可以讓類比內容可定址記憶體裝置300在進行記憶體內部之資料搜尋與比對,以判斷記憶胞所儲存的資料以及所輸入的資料是否匹配時,可以讓導通的記憶胞以固定的電流來使匹配訊號線312的電壓下降,讓匹配訊號線312的電壓下降的時間係為可控制的,以減少誤判的機率。By using the memory cell shown in FIG. 4, the analog content
請參照第8圖,其繪示根據本揭露另一實施例之類比內容可定址記憶體的記憶胞400之電路圖。記憶胞400包括N型電晶體402、P型電晶體404、電流控制電路406與及邏輯閘408,N型電晶體402具有閘極G1,N型電晶體402的閘極G1用以接收輸入訊號B(i)。P型電晶體404具有閘極G2,P型電晶體404的閘極G2用以接收輸入訊號A(i)。電流控制電路406係具有第一感測放大器406A與第二感測放大器406B。及邏輯閘408係耦接至第一感測放大器406A與第二感測放大器406B,第一感測放大器406A與第二感測放大器406B各具有一臨界電流值。N型電晶體402係與第一感測放大器406A電性連接,P型電晶體404係與第二感測放大器406B電性連接。當輸入訊號B(i)的輸入電壓與輸入訊號A(i)的輸入電壓位於匹配範圍內時,N型電晶體402對應之至少一導通電流之一I1與P型電晶體404對應之至少一導通電流之另一I2均大於或等於臨界電流值,及邏輯閘408輸出一第一邏輯值。Please refer to FIG. 8 , which shows a circuit diagram of a
請同時參考第8圖及第9圖,其中第9圖繪示具有電流控制電路406之記憶胞400的匹配範圍示意圖。匹配範圍係與N型電晶體402的臨界電壓Vthn、P型電晶體404的臨界電壓Vthp、及第一感測放大器406A與第二感測放大器406B的臨界電流值相關。如第9圖所示,當第一感測放大器406A與第二感測放大器406B的臨界電流值為臨界電流值Ith1時,匹配範圍MRa為N型電晶體402之導通電流為臨界電流值Ith1的閘極電壓V1’與P型電晶體404的之導通電流為臨界電流值Ith1的閘極電壓V6’所決定。當第一感測放大器406A與第二感測放大器406B的臨界電流值為臨界電流值Ith2時,匹配範圍MRa為N型電晶體402之導通電流為臨界電流值Ith2的閘極電壓V2’與P型電晶體404的之導通電流為臨界電流值Ith1的閘極電壓V5’所決定。當第一感測放大器406A與第二感測放大器406B的臨界電流值為臨界電流值Ith3時,匹配範圍MRa為N型電晶體402之導通電流為臨界電流值Ith3的閘極電壓V3’與P型電晶體404的之導通電流為臨界電流值Ith1的閘極電壓V4’所決定。Please refer to FIG. 8 and FIG. 9 at the same time, wherein FIG. 9 shows a schematic diagram of the matching range of the
茲以第一感測放大器406A與第二感測放大器406B之臨界電流值為臨界電流值Ith1為例做說明。請同時參考第8圖與第9圖。當輸入訊號A(i)的輸入電壓與輸入訊號B(i)的輸入電壓位於匹配範圍MRa內時,N型電晶體402產生第一電流I1,P型電晶體404產生第二電流I2。當第一電流I1及第二電流I2均大於或等於第一感測放大器406A與第二感測放大器406B的臨界電流值,第一感測放大器406A與第二感測放大器406B將輸出第一邏輯值,例如是邏輯值1。當及邏輯閘408的兩個輸入端的輸入均為第一邏輯值時,及邏輯閘408將輸出第一邏輯值(例如是邏輯值1),以指示輸入訊號A(i)的輸入電壓與輸入訊號B(i)的輸入電壓均位於匹配範圍MRa內。The critical current value Ith1 of the
因此,依據第8圖的實施例,記憶胞400的第一感測放大器406A與第二感測放大器406B可分別於輸入訊號A(i)與輸入訊號B(i)位於匹配範圍內時,以相同的第一感測放大器406A與第二感測放大器406B的臨界電流值來進行N型電晶體402與P型電晶體404是否導通的判斷,而可達到減少誤判之機率的功效。Therefore, according to the embodiment shown in FIG. 8, the
請參照第10圖,其繪示應用第8圖之記憶胞之類比內容可定址記憶體裝置500的電路圖。類比內容可定址記憶體裝置500包括第一字元線驅動電路502、第二字元線驅動電路504、第一N通道NAND串列組506、第一P通道NAND串列組508、多個第一感測放大電路510(1), 510(2), …, 510(n)、多個第二感測放大電路512(1), 512(2), …, 512(n)與多個第一及邏輯閘514(1), 514(2), …, 514(n)。第一字元線驅動電路502用以提供多個第一輸入訊號SL1(1), SL1(2), …, SL1(m),第二字元線驅動電路504用以提供多個第二輸入訊號SL2(1), SL2(2), …, SL2(m)。第一N通道NAND串列組506包括多個第一N通道NAND串列516(1), 516(2), …, 516(n),各第一N通道NAND串列516用以接收第一輸入訊號SL1(1), SL1(2), …, SL1(m),各第一N通道NAND串列516更用以產生第一電流,例如第一N通道NAND串列516(1), 516(2), …, 516(n)分別產生第一電流I1(1), I1(2), …, I1(n)。第一P通道NAND串列組508包括多個第一P通道NAND串列518(1), 518(2), …, 518(n),各第一P通道NAND串列518用以接收第二輸入訊號SL2(1), SL2(2), …, SL2(m)。各第一P通道NAND串列518更用以產生一第二電流,例如第一P通道NAND串列518(1), 518(2), …, 518(n)分別產生第二電流I2(1), I2(2), …, I2(n)。多個第一感測放大電路510(1), 510(2), …, 510(n)係分別耦接至第一N通道NAND串列組506之第一N通道NAND串列516(1), 516(2), …, 516(n)。第二感測放大電路512(1), 512(2), …, 512(n)係分別耦接至第一P通道NAND串列組508之多個第一P通道NAND串列518(1), 518(2), …, 518(n),第一感測放大電路510(1), 510(2), …, 510(n)與第二感測放大電路512(1), 512(2), …, 512(n)係各具有一臨界電流值。各個第一及邏輯閘係耦接至對應之第一感測放大電路510與對應之第二感測放大電路512。其中第一及邏輯閘514(1), 514(2), …, 514(n)之一為選定之第一及邏輯閘514(i)(i為1至n中之一正整數)。當選定之第一及邏輯閘510(i)所對應之第一電流I1(i)與第二電流I2(i)的電流均大於或等於臨界電流值時,選定之第一及邏輯閘514(i)輸出一第一邏輯值至解碼器520。此時解碼器520將判斷出第一字元線驅動電路502所提供之資料係與第一N通道NAND串列516(i)所儲存的資料匹配,第二字元線驅動電路504所提供之資料係與第一P通道NAND串列518(i)所儲存的資料匹配。Please refer to FIG. 10 , which shows a circuit diagram of an analog content
因此,依據第9圖,類比內容可定址記憶體裝置500的第一感測放大電路510(i)與第二感測放大電路512(i)可分別於第一電流I1(i)與第二電流I2(i)大於或等於相同的電流位準,亦即是第一感測放大電路510(i)與第二感測放大電路512(i)的臨界電流值時,即可讓第一感測放大電路510(i)與第二感測放大電路512(i)輸出第一邏輯值。如此,可以避免於進行記憶體內部之類比資料的搜尋與比對時,因為第一電流I1(i)或第二電流I2(i)的大小不同,而使得偵測是否匹配所需的時間長度有所差異,而造成誤判的情況。Therefore, according to FIG. 9, the first sense amplifier circuit 510(i) and the second sense amplifier circuit 512(i) of the analog content
更進一步來說,如第10圖所示,第一N通道NAND串列組506之第一N通道NAND串列516(i)的一端用以接收第一位元線訊號BL1(i),第一N通道NAND串列組506之第一N通道NAND串列516(i)的另一端則耦接至對應之第一感測放大電路510(i)。第一P通道NAND串列組508之第一P通道NAND串列518(i)的一端用以接收第二位元線訊號BL2(i),第一P通道NAND串列組508之第一P通道NAND串列518(i)的另一端則是耦接至第二感測放大電路512(i)。第一N通道NAND串列組506之第一N通道NAND串列516(i)包括多個第一N型電晶體,第一P通道NAND串列組508之第一P通道NAND串列518(i)包括多個第一P型電晶體。第一N通道NAND串列516(i)之第一N型電晶體的閘極用以接收第一輸入訊號SL1(1), SL1(2), …, SL1(m)。第一P通道NAND串列518(i)之第一P型電晶體的閘極用以接收第二輸入訊號SL2(1), SL2(2), …, SL2(m)。Furthermore, as shown in FIG. 10, one end of the first N-channel NAND series 516(i) of the first N-channel
假設多個第一N通道NAND串列之一(例如是第一N通道NAND串列516(i),i為1至n之正整數)所儲存的資料範圍與輸入訊號匹配。當第一字元線驅動電路502所提供之第一輸入訊號SL1(1), SL1(2), …, SL1(m)與第一N通道NAND串列516(i)中之m個記憶胞的儲存範圍匹配時,亦即是第一輸入訊號SL1(1), SL1(2), …, SL1(m)的電壓位準分別位於第一N通道NAND串列516(i)中之m個記憶胞的匹配範圍內時,則第一N通道NAND串列516(i)將輸出第一電流I1(i)。同理,當第二字元線驅動電路504所提供之第二輸入訊號SL2(1), SL2(2), …, SL2(m)與第一P通道NAND串列518(i)中之m個記憶胞的儲存範圍匹配時,亦即是第二輸入訊號SL2(1), SL2(2), …, SL2(m)的電壓位準分別位於第一P通道NAND串列518(i)中之m個記憶胞的匹配範圍內時,則第一P通道NAND串列518(i)將輸出第二電流I2(i)。此時,當第一電流I1(i)與第二電流I2(i)的電流均大於或等於臨界電流值時,第一及邏輯閘514(i)輸出一第一邏輯值,代表第一字元線驅動電路502所提供之資料係與第一N通道NAND串列516(i)所儲存的資料匹配,且第二字元線驅動電路504所提供之資料係與第一P通道NAND串列518(i)所儲存的資料匹配。Assume that the range of data stored in one of the first N-channel NAND series (for example, the first N-channel NAND series 516(i), where i is a positive integer from 1 to n) matches the input signal. When the first input signal SL1(1), SL1(2), . When the storage ranges match, that is, the voltage levels of the first input signals SL1(1), SL1(2), . When the memory cell is within the matching range, the first N-channel NAND series 516(i) will output the first current I1(i). Similarly, when the second input signal SL2(1), SL2(2), . When the storage ranges of the memory cells match, that is, the voltage levels of the second input signals SL2(1), SL2(2), ..., SL2(m) are respectively located in the first P-channel NAND series 518(i) When the m memory cells are within the matching range, the first P-channel NAND series 518(i) will output the second current I2(i). At this time, when the currents of the first current I1(i) and the second current I2(i) are greater than or equal to the critical current value, the first logic gate 514(i) outputs a first logic value, representing the first word The data provided by the word
在一實施例中,第一N通道NAND串列組506與第一P通道NAND串列組508為2D快閃記憶體結構或3D快閃記憶體結構。在一實施例中,第一N通道NAND串列組506與第一P通道NAND串列組508可以使用F-N穿隧、通道熱電子注入技術或是多晶矽對多晶矽的方式來進行編程。在一實施例中,第一N通道NAND串列組506與第一P通道NAND串列組508可以使用F-N穿隧、帶對帶熱電洞入射或是多晶矽對多晶矽的方式來進行抹除。In one embodiment, the first N-channel NAND series set 506 and the first P-channel NAND series set 508 are 2D flash memory structure or 3D flash memory structure. In one embodiment, the first N-
在一實施例中,類比內容可定址記憶體裝置500係為快閃記憶體,快閃記憶體為電荷儲存記憶體、電荷捕捉記憶體、分離式閘極記憶體或鐵電體場效電晶體記憶體。在另一實施例中,類比內容可定址記憶體裝置500係為超陡峭快閃記憶體,超陡峭快閃記憶體為閘流體隨機存取記憶體、閘極控制閘流體、穿隧式場效電晶體、或負電容場效電晶體。In one embodiment, the analog content
請參照第11圖,其繪示根據另一實施例之類比內容可定址記憶體裝置600。類比內容可定址記憶體裝置600包括第10圖的類比內容可定址記憶體裝置500、第三字元線驅動電路602、第四字元線驅動電路604、第二N通道NAND串列組606、第二P通道NAND串列組608、多個第三感測放大電路610(1), 610(2), …, 610(n)、多個第四感測放大電路612(1), 612(2), …, 612(n)、多個第二及邏輯閘614(1), 614(2), …, 614(n)與多個第三及邏輯閘622(1), 622(2), …, 622(n)。第三字元線驅動電路602用以提供多個第三輸入訊號SL3(1), SL3(2), …, SL3(m),第四字元線驅動電路604用以提供多個第四輸入訊號SL4(1), SL4(2), …, SL4(m)。第二N通道NAND串列組606包括多個第二N通道NAND串列616(1), 616(2), …, 616(n),各第二N通道NAND串列用以接收第三輸入訊號SL3(1), SL3(2), …, SL3(m)。各第二N通道NAND串列更用以產生一第三電流,例如第二N通道NAND串列616(1), 616(2), …, 616(n)分別產生第三電流I3(1), I3(2), …, I3(n)。第二P通道NAND串列組608包括多個第二P通道NAND串列618(1), 618(2), …, 618(n)。各第二P通道NAND串列用以接收第四輸入訊號SL4(1), SL4(2), …, SL4(m),各第二P通道NAND串列更用以產生一第四電流,例如第二P通道NAND串列618(1), 618(2), …, 618(n)分別產生第四電流I4(1), I4(2), …, I4(n)。第三感測放大電路610(1), 610(2), …, 610(n)係耦接至第二N通道NAND串列組606之多個第二N通道NAND串列616(1),616(2), …, 616(n),多個第四感測放大電路612(1), 612(2), …, 612(n)係耦接至第二P通道NAND串列組608之多個第二P通道NAND串列618(1), 618(2), …, 618(n),多個第三感測放大電路610(1), 610(2), …, 610(n)與多個第四感測放大電路612(1), 612(2), …, 612(n)係各具有一臨界電流值。各個第二及邏輯閘614(1), 614(2), …, 614(n)係耦接至對應之第三感測放大電路與對應之第四感測放大電路。各個第三及邏輯閘622(1), 622(2), …, 622(n)係耦接至對應之第一及邏輯閘及對應之第二及邏輯閘。其中,第二及邏輯閘614(1), 614(2), …, 614(n)之一為選定之第二及邏輯閘614(i), i為1至n中之一正整數。當選定之第二及邏輯閘614(i)所對應之第三電流I3(i)與第四電流I4(i)的電流均大於或等於臨界電流值時,選定之第二及邏輯閘輸614(i)出第一邏輯值。第三及邏輯閘622(1), 622(2), …, 622(n)之一為選定之第三及邏輯閘622(i)。當選定之第三及邏輯閘622(i)所對應之第一及邏輯閘514(i)與第二及邏輯閘614(i)均輸出第一邏輯值時,選定之第三及邏輯閘輸622(i)出第一邏輯值。Please refer to FIG. 11 , which shows an analog content
如第11圖所示,第二N通道NAND串列組606之各個第二N通道NAND串列616(1), 616(2), …, 616(n)的一端用以接收一第三位元線訊號,例如第二N通道NAND串列616(1), 616(2), …, 616(n)分別接收第三位元線訊號BL3(1), BL3(2), …, BL3(n)。第二N通道NAND串列組606之各個第二N通道NAND串列616(1), 616(2), …, 616(n)的另一端耦接至對應之第三感測放大電路,例如各個第二N通道NAND串列616(1), 616(2), …, 616(n)分別耦接至第三感測放大電路610(1), 610(2), …, 610(n)。第二P通道NAND串列組608之各個第二P通道NAND串列618(1), 618(2), …, 618(n)的一端用以接收第四位元線訊號,例如第二P通道NAND串列618(1), 618(2), …, 618(n)接收第四位元線訊號BL4(1), BL4(2), …, BL4(n)。第二P通道NAND串列組608之各個第二P通道NAND串列618(1), 618(2), …, 618(n)的另一端耦接至對應之第四感測放大電路,例如第二P通道NAND串列618(1), 618(2), …, 618(n)的另一端耦接至第四感測放大電路612(1), 612(2), …, 612(n)。第二N通道NAND串列組606之各個第二N通道NAND串列包括多個第二N型電晶體。第二P通道NAND串列組608之各個第二P通道NAND串列包括多個第二P型電晶體。各個第二N通道NAND串列之各個第二N型電晶體的閘極用以接收對應之第三輸入訊號。各個第二P通道NAND串列之各個第二P型電晶體的閘極用以接收對應之第四輸入訊號。因此,依據第11圖的實施例,類比內容可定址記憶體裝置600的第一感測放大電路510(1), 510(2), …, 510(n)、第二感測放大電路512(1), 512(2), …, 512(n)、第三感測放大電路610(1), 610(2), …, 610(n)與第四感測放大電路612(1), 612(2), …, 612(n)係分別對應第一輸入訊號SL1(1), SL1(2), …, SL1(m)、第二輸入訊號SL2(1), SL2(2), …, SL2(m)、第三輸入訊號SL3(1), SL3(2), …, SL3(m)與第四輸入訊號SL4(1), SL4(2), …, SL4(m)。當第三字元線驅動電路602所提供之第三輸入訊號SL3(1), SL3(2), …, SL3(m)與第二N通道NAND串列616(i)中之m個記憶胞的儲存範圍匹配時,亦即是第三輸入訊號SL3(1), SL3(2), …, SL3(m)的電壓位準分別位於第二N通道NAND串列616(i)中之m個記憶胞的匹配範圍內時,則第二N通道NAND串列616(i)將輸出第三電流I3(i)。同理,當第四字元線驅動電路604所提供之第四輸入訊號SL4(1), SL4(2), …, SL4(m)與第二P通道NAND串列618(i)中之m個記憶胞的儲存範圍匹配時,亦即是第四輸入訊號SL4(1), SL4(2), …, SL4(m)的電壓位準分別位於第二P通道NAND串列618(i)中之m個記憶胞的匹配範圍內時,則第二P通道NAND串列618(i)將輸出第四電流I4(i)。此時,當第三電流I3(i)與第四電流I4(i)的電流均大於或等於臨界電流值時,第二及邏輯閘614(i)輸出一第一邏輯值,代表第三字元線驅動電路602所提供之資料係與第二N通道NAND串列616(i)所儲存的資料匹配,且第四字元線驅動電路604所提供之資料係與第二P通道NAND串列618(i)所儲存的資料匹配。As shown in FIG. 11, one end of each second N-channel NAND series 616(1), 616(2), ..., 616(n) of the second N-channel
當第一及邏輯閘514(i)輸出第一邏輯值,且第二及邏輯閘614(i)輸出第一邏輯值時,第三及邏輯閘622(i)將輸出第一邏輯值。如第11圖所示,當i=2時,則第三及邏輯閘622(2)將輸出第一邏輯值(例如邏輯值“1”)至解碼器620。此時解碼器620將判斷出第一字元線驅動電路502所提供之資料係與第一N通道NAND串列516(i)所儲存的資料匹配,第二字元線驅動電路504所提供之資料係與第一P通道NAND串列518(i)所儲存的資料匹配,第三字元線驅動電路602所提供之資料係與第二N通道NAND串列616(i)所儲存的資料匹配,且第四字元線驅動電路604所提供之資料係與第二P通道NAND串列618(i)所儲存的資料匹配。如此,透過類比內容可定址記憶體裝置600,可增加搜尋的資料量(即增加輸入訊號量),例如可以增加第三字元線驅動電路602所提供之資料以及第四字元線驅動電路604所提供之資料來進行資料搜尋與比對,以提高類比內容可定址記憶體的效能。再者,類比內容可定址記憶體裝置600更可減少N通道NAND串列與P通道NAND串列的大小,以降低RC延遲,以加速類比內容可定址記憶體裝置的反應速度。When the first sum logic gate 514(i) outputs the first logic value and the second sum logic gate 614(i) outputs the first logic value, the third sum logic gate 622(i) will output the first logic value. As shown in FIG. 11 , when i=2, the third AND logic gate 622 ( 2 ) will output the first logic value (eg logic value “1”) to the decoder 620 . At this time, the decoder 620 will judge that the data provided by the first word
根據上述之實施例,本案提出之類比內容可定址記憶體的記憶胞以及類比內容可定址記憶體裝置利用電流控制電路來固定匹配範圍內之不同輸入訊號的電流位準,使匹配訊號線的放電時間穩定,以減少資料搜尋及比對時的誤判機率,並可加快類比內容可定址記憶體裝置的處理速度。According to the above-mentioned embodiment, the memory cell of the analog content addressable memory and the analog content addressable memory device proposed in this case use the current control circuit to fix the current level of different input signals within the matching range, so that the discharge of the matching signal line The time is stable to reduce the probability of misjudgment during data search and comparison, and to speed up the processing speed of analog content addressable memory devices.
綜上所述,雖然本發明已以實施例揭露如上,然其並非用以限定本發明。本發明所屬技術領域中具有通常知識者,在不脫離本發明之精神和範圍內,當可作各種之更動與潤飾。因此,本發明之保護範圍當視後附之申請專利範圍所界定者為準。To sum up, although the present invention has been disclosed by the above embodiments, it is not intended to limit the present invention. Those skilled in the art of the present invention can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be defined by the scope of the appended patent application.
100,200,400:類比內容可定址記憶體 202,304:記憶胞 204,312,ML:匹配訊號線 206,306,402:N型電晶體 208,308,404:P型電晶體 210,310,406:電流控制電路 210A,310A:控制電晶體 300,500,600:類比內容可定址記憶體裝置 302:字元線驅動電路 314,SL:源極線 316:源極線驅動電路 318:感測放大器電路 406A:第一感測放大器 406B:第二感測放大器 408:及邏輯閘 502:第一字元線驅動電路 504:第二字元線驅動電路 506:第一N通道NAND串列組 508:第一P通道NAND串列組 510(1),510(2),510(n):第一感測放大電路 512(1),512(2),512(n):第二感測放大電路 514(1),514(2),514(n):第一及邏輯閘 516(1), 516(2),516(n):第一N通道NAND串列 518(1),518(2),518(n):第一P通道NAND串列 520,620:解碼器 602:第三字元線驅動電路 604:第四字元線驅動電路 606:第二N通道NAND串列組 608:第二P通道NAND串列組 610(1),610(2),610(n):第三感測放大電路 612(1),612(2),612(n):第四感測放大電路 614(1),614(2),614(n):第二及邏輯閘 616(1),616(2),616(n):第二N通道NAND串列 618(1),618(2),618(n):第二P通道NAND串列 622(1),622(2),622(n):第三及邏輯閘S1,A(1),A(2),A(n),B(1),B(2),B(n):輸入訊號 CL1:類比CAM記憶胞 R(1),R(2),R(3),R(4):列 Rn:匹配不成功結果 Ry:匹配成功結果 MSn:第一浮動閘極裝置 MSp:第二浮動閘極裝置 Ip,I pass:導通電流 MR,MR0,MR1,MRa,MRb,MRc:匹配範圍 CN,CP:曲線 LB:下限 UB:上限 G1,G2,G3:閘極 D1,D2,D3:汲極 S1,S2,S3:源極 C(1),C(2),…,C(n),C(i):控制電壓 I1(1),I1(2),I1(n):第一電流 I2(1),I2(2),I2(n):第二電流 I3(1),I3(2),I3(n):第三電流 I4(1),I4(2),I4(n):第四電流 SL1(1),SL1(2),SL1(m):第一輸入訊號 SL2(1),SL2(2),SL2(m):第二輸入訊號 SL3(1),SL3(2),SL3(m):第三輸入訊號 SL4(1),SL4(2),SL4(m):第四輸入訊號 BL1(1),BL1(2),BL1(3),BL1(n):第一位元線訊號 BL2(1),BL2(2),BL2(3),BL2(n):第二位元線訊號 BL3(1),BL3(2),BL3(3),BL3(n):第三位元線訊號 BL4(1),BL4(2),BL4(3),BL4(n):第四位元線訊號100, 200, 400: analog content addressable memory 202, 304: memory cell 204, 312, ML: matching signal line 206, 306, 402: N-type transistor 208, 308, 404: P-type transistor 210, 310, 406: current control circuit 210A, 310A: control transistor 300, 500, 600: analog content addressable memory Bulk device 302: word line drive circuit 314, SL: source line 316: source line drive circuit 318: sense amplifier circuit 406A: first sense amplifier 406B: second sense amplifier 408: and logic gate 502: First word line driving circuit 504: second word line driving circuit 506: first N-channel NAND string group 508: first P-channel NAND string group 510(1), 510(2), 510(n) : first sense amplifier circuits 512(1), 512(2), 512(n): second sense amplifier circuits 514(1), 514(2), 514(n): first and logic gates 516( 1), 516(2), 516(n): first N-channel NAND series 518(1), 518(2), 518(n): first P-channel NAND series 520, 620: decoder 602: third Word line driving circuit 604: fourth word line driving circuit 606: second N-channel NAND string group 608: second P-channel NAND string group 610(1), 610(2), 610(n): the first Three sense amplifier circuits 612(1), 612(2), 612(n): fourth sense amplifier circuits 614(1), 614(2), 614(n): second and logic gates 616(1) , 616(2), 616(n): the second N-channel NAND series 618(1), 618(2), 618(n): the second P-channel NAND series 622(1), 622(2), 622(n): third logic gate S1, A(1), A(2), A(n), B(1), B(2), B(n): input signal CL1: analog CAM memory cell R(1), R(2), R(3), R(4): column Rn: matching unsuccessful result Ry: matching successful result MSn: first floating gate device MSp: second floating gate device Ip, I pass : conduction current MR, MR0, MR1, MRa, MRb, MRc: matching range CN, CP: curve LB: lower limit UB: upper limit G1, G2, G3: gate D1, D2, D3: drain S1, S2, S3: source C(1), C(2),..., C(n), C(i): control voltage I1(1), I1(2), I1(n): first current I2(1) , I2(2), I2(n): the second current I3(1), I3(2), I3(n): the third current I4(1), I4(2), I4(n): the fourth current SL1(1), SL1(2 ), SL1(m): the first input signal SL2(1), SL2(2), SL2(m): the second input signal SL3(1), SL3(2), SL3(m): the third input signal SL4 (1), SL4(2), SL4(m): the fourth input signal BL1(1), BL1(2), BL1(3), BL1(n): the first bit line signal BL2(1), BL2 (2), BL2(3), BL2(n): second bit line signal BL3(1), BL3(2), BL3(3), BL3(n): third bit line signal BL4(1) , BL4(2), BL4(3), BL4(n): the fourth bit line signal
第1圖繪示類比內容可定址記憶體(analog content-addressable memory, analog CAM)之操作示意圖;
第2圖繪示類比CAM記憶胞;
第3圖繪示繪示類比CAM記憶胞的匹配範圍;
第4圖繪示根據一實施例之類比內容可定址記憶體的記憶胞之電路圖;
第5A圖繪示未具有電流控制電路之記憶胞的匹配範圍;
第5B圖繪示連接於第5A圖的記憶胞之匹配訊號線的輸出電壓與放電時間的關係圖;
第6A圖繪示具有電流控制電路之記憶胞的匹配範圍;
第6B圖繪示連接於第6A圖的記憶胞之匹配訊號線的輸出電壓與放電時間的關係圖;
第7圖繪示將第4圖所示之記憶胞應用於類比內容可定址記憶體裝置300之一例;
第8圖繪示根據另一實施例之類比內容可定址記憶體的記憶胞之電路圖;
第9圖繪示繪示具有電流控制電路之記憶胞的匹配範圍示意圖;
第10圖繪示應用第8圖之記憶胞之類比內容可定址記憶體裝置的電路圖。
第11圖繪示根據另一實施例之類比內容可定址記憶體裝置。
Figure 1 shows a schematic diagram of the operation of an analog content-addressable memory (analog CAM);
Figure 2 shows an analog CAM memory cell;
Figure 3 shows the matching range of the analog CAM memory cell;
FIG. 4 shows a circuit diagram of a memory cell of an analog content addressable memory according to an embodiment;
Figure 5A shows the matching range of a memory cell without a current control circuit;
Figure 5B shows the relationship between the output voltage and the discharge time of the matching signal line connected to the memory cell in Figure 5A;
Figure 6A shows the matching range of a memory cell with a current control circuit;
Figure 6B shows the relationship between the output voltage and the discharge time of the matching signal line connected to the memory cell in Figure 6A;
FIG. 7 shows an example of applying the memory cell shown in FIG. 4 to an analog content
200:類比內容可定址記憶體 200: Analog Content Addressable Memory
202:記憶胞 202: memory cell
204:匹配訊號線 204: Matching signal line
206:N型電晶體 206: N-type transistor
208:P型電晶體 208: P-type transistor
210:電流控制電路 210: current control circuit
210A:控制電晶體 210A: control transistor
A(i),B(i):輸入訊號 A(i), B(i): input signal
C(i):控制電壓 C(i): control voltage
G1,G2,G3:閘極 G1, G2, G3: gate
D1,D2,D3:汲極 D1, D2, D3: drain
S1,S2,S3:源極 S1, S2, S3: source
Ipass:導通電流 I pass : pass current
SL:源極線 SL: source line
Claims (20)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW110132411A TWI788987B (en) | 2021-09-01 | 2021-09-01 | Memory cell for analog content-addressable memory and analog content-addressable memory device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW110132411A TWI788987B (en) | 2021-09-01 | 2021-09-01 | Memory cell for analog content-addressable memory and analog content-addressable memory device |
Publications (2)
Publication Number | Publication Date |
---|---|
TWI788987B true TWI788987B (en) | 2023-01-01 |
TW202312169A TW202312169A (en) | 2023-03-16 |
Family
ID=86669934
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW110132411A TWI788987B (en) | 2021-09-01 | 2021-09-01 | Memory cell for analog content-addressable memory and analog content-addressable memory device |
Country Status (1)
Country | Link |
---|---|
TW (1) | TWI788987B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6256216B1 (en) * | 2000-05-18 | 2001-07-03 | Integrated Device Technology, Inc. | Cam array with minimum cell size |
US6317349B1 (en) * | 1999-04-16 | 2001-11-13 | Sandisk Corporation | Non-volatile content addressable memory |
US6339540B1 (en) * | 2000-12-05 | 2002-01-15 | Tower Semiconductor Ltd. | Content-addressable memory for virtual ground flash architectures |
US7177183B2 (en) * | 2003-09-30 | 2007-02-13 | Sandisk 3D Llc | Multiple twin cell non-volatile memory array and logic block structure and method therefor |
US20090190404A1 (en) * | 2008-01-25 | 2009-07-30 | Roohparvar Frankie F | Nand flash content addressable memory |
-
2021
- 2021-09-01 TW TW110132411A patent/TWI788987B/en active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6317349B1 (en) * | 1999-04-16 | 2001-11-13 | Sandisk Corporation | Non-volatile content addressable memory |
US6256216B1 (en) * | 2000-05-18 | 2001-07-03 | Integrated Device Technology, Inc. | Cam array with minimum cell size |
US6339540B1 (en) * | 2000-12-05 | 2002-01-15 | Tower Semiconductor Ltd. | Content-addressable memory for virtual ground flash architectures |
US7177183B2 (en) * | 2003-09-30 | 2007-02-13 | Sandisk 3D Llc | Multiple twin cell non-volatile memory array and logic block structure and method therefor |
US20090190404A1 (en) * | 2008-01-25 | 2009-07-30 | Roohparvar Frankie F | Nand flash content addressable memory |
Also Published As
Publication number | Publication date |
---|---|
TW202312169A (en) | 2023-03-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7391652B2 (en) | Method of programming and erasing a p-channel BE-SONOS NAND flash memory | |
KR960003398B1 (en) | Nonvolatile semiconductor memory device having row decoder supplying a negative potential to word lines during erasing mode | |
US7262995B2 (en) | Method for reading flash memory cell, NAND-type flash memory apparatus, and NOR-type flash memory apparatus | |
US9147485B2 (en) | Memory page buffer | |
KR101844074B1 (en) | Architecture for a 3d memory array | |
US7751243B2 (en) | Semiconductor memory device provided with MOS transistor having charge accumulation layer and control gate and data write method of NAND flash memory | |
US20140198576A1 (en) | Programming technique for reducing program disturb in stacked memory structures | |
US10803952B2 (en) | Vertical memory device having improved electrical characteristics and method of operating the same | |
KR101556681B1 (en) | Method for reading of the cell string | |
US8842475B2 (en) | Configuration memory | |
JP2010198685A (en) | Nonvolatile semiconductor memory | |
US20230368842A1 (en) | Content-addressable memory and analog content-addressable memory device | |
US10658057B2 (en) | Semiconductor memory device | |
US11328775B2 (en) | Ternary content addressable memory and memory cell thereof | |
CN115394339A (en) | Analog content addressable memory and method of operating the same | |
TWI788987B (en) | Memory cell for analog content-addressable memory and analog content-addressable memory device | |
US6141244A (en) | Multi level sensing of NAND memory cells by external bias current | |
KR20090039063A (en) | Non-volatile memory device | |
TWI763493B (en) | Analog content-address memory and operation method thereof | |
US11670373B2 (en) | Three-dimensional memory device programming with reduced threshold voltage shift | |
US11205492B2 (en) | Responding to power loss | |
TWI752667B (en) | Ternary content addressable memory and memory cell thereof | |
US6778439B2 (en) | Nonvolatile semiconductor memory device with MONOS type memory cell | |
TWI489593B (en) | Hot carrier programming of nand flash memory | |
Ohnakado et al. | Novel self-limiting high-speed program scheme of p-channel DINOR flash memory with n-channel select transistors |