TWI874275B - Memory module with improved timing adaptivity of sensing amplification - Google Patents
Memory module with improved timing adaptivity of sensing amplification Download PDFInfo
- Publication number
- TWI874275B TWI874275B TW113132059A TW113132059A TWI874275B TW I874275 B TWI874275 B TW I874275B TW 113132059 A TW113132059 A TW 113132059A TW 113132059 A TW113132059 A TW 113132059A TW I874275 B TWI874275 B TW I874275B
- Authority
- TW
- Taiwan
- Prior art keywords
- voltage
- node
- memory module
- pulse width
- input terminal
- Prior art date
Links
Images
Landscapes
- Static Random-Access Memory (AREA)
Abstract
Description
本發明係關於一種可改良感測放大時序適應性的記憶模組,且特別是關於一種可使感測放大器致能之時間長短適應輸出入數目及/或感測放大器供應電壓之記憶模組。The present invention relates to a memory module capable of improving the timing adaptability of a sense amplifier, and in particular to a memory module capable of adapting the duration of enabling a sense amplifier to the input/output number and/or the supply voltage of the sense amplifier.
記憶模組,例如嵌入式靜態隨機存取記憶模組,是積體電路(半導體晶片)的重要基礎構築方塊。Memory modules, such as embedded static random access memory modules, are important basic building blocks of integrated circuits (semiconductor chips).
記憶模組中設有多個記憶單元與複數個感測放大器;這些感測放大器受控於一致能信號。當致能信號被激發時,各感測放大器會致能而運作;當致能信號被停止激發時,各感測放大器便會失能而不運作。The memory module is provided with a plurality of memory cells and a plurality of sense amplifiers; these sense amplifiers are controlled by an enable signal. When the enable signal is activated, each sense amplifier is activated and operates; when the enable signal is stopped, each sense amplifier is disabled and does not operate.
當記憶模組要在一讀取回合(read cycle)中讀出一記憶單元中儲存的資料時,該記憶單元對應的位元線會被導通至一對應的感測放大器,然後致能信號會被激發,使該感測放大器致能以感測該對應位元線的電壓,並據以判斷該記憶單元儲存的資料;之後,致能信號會被停止激發,使該感測放大器失能而停止感測,該讀取回合也就此結束。在各讀取回合中,感測放大器致能運作的時間長短取決於致能信號被激發的時間長短,也就是致能信號的脈寬。若致能信號的脈寬不足,感測放大器致能運作的時間就會不足,資料讀取的正確性也會連帶受影響。When the memory module wants to read the data stored in a memory cell in a read cycle, the bit line corresponding to the memory cell will be connected to a corresponding sense amplifier, and then the enable signal will be activated, enabling the sense amplifier to sense the voltage of the corresponding bit line and judge the data stored in the memory cell accordingly; after that, the enable signal will be stopped, disabling the sense amplifier and stopping sensing, and the read cycle will end. In each read cycle, the length of time the sense amplifier is enabled depends on the length of time the enable signal is activated, that is, the pulse width of the enable signal. If the pulse width of the enable signal is insufficient, the sense amplifier will not have enough time to operate, and the accuracy of data reading will also be affected.
在習知技術中,習知記憶模組會設置固定數目個串接的反相器,致能信號的脈寬就取決於這些串接反相器的閘延遲(gate delay)總和。因此,在習知技術中,致能信號的脈寬缺乏彈性,難以適應各種記憶模組的不同需求。並且,這些串接反相器也會佔用不少佈局面積。In the known technology, a fixed number of inverters connected in series are set in the known memory module, and the pulse width of the enable signal depends on the sum of the gate delays of these inverters connected in series. Therefore, in the known technology, the pulse width of the enable signal lacks flexibility and is difficult to adapt to the different requirements of various memory modules. In addition, these inverters connected in series will also occupy a lot of layout area.
本發明的目的之一是提供一種可改良感測放大時序適應性的記憶模組(例如100,圖1),包含至少一位元線(例如BL[q])、至少一字元線(例如WL[p])、一追隨位元線(例如TBL)、一追隨字元線(例如TWL)、至少一記憶單元(例如c[p,q])、至少一追隨單元(例如tc[p])、至少一感測放大器(例如SA[k])與一脈寬控制器(例如300,圖4)。該追隨字元線包含一前節點(例如w1)與一後節點(例如w3),且於該前節點與該後節點間的長度係正相關於各該字元線的長度。各該記憶單元耦接該至少一位元線的其中之一以及該至少一字元線的其中之一。各該追隨單元耦接該追隨位元線。各該感測放大器耦接該至少一位元線的其中之一,並接收一致能信號(例如GS);各該感測放大器在該致能信號被激發(activated)時受控致能,並在該致能信號被停止激發時失能。脈寬控制器耦接於該追隨位元線、該前節點與該後節點,並提供該致能信號。其中,當該追隨位元線的電壓(例如vTBL,圖5)改變至一預設電壓時(例如vt0,在時點t3,圖5),該脈寬控制器激發該致能信號(例如在時點t6,圖5),並使該前節點的電壓(例如vw1,圖5)改變(例如在時點t7開始由電壓v4切換至電壓v3,圖5)。當該前節點的電壓改變,該追隨字元線在一第一延遲時間(例如d1,圖5)後使該後節點的電壓(例如vw3,圖5)改變(例如在時點t8開始由電壓v4切換至電壓v3,圖5)。當該後節點的電壓改變,該脈寬控制器在一第二延遲時間(例如d2,圖5)後停止激發該致能信號(例如在時點t12,圖5)。One of the purposes of the present invention is to provide a memory module (e.g., 100, FIG. 1 ) capable of improving the adaptability of sense amplification timing, comprising at least one bit line (e.g., BL[q]), at least one word line (e.g., WL[p]), a tracking bit line (e.g., TBL), a tracking word line (e.g., TWL), at least one memory cell (e.g., c[p,q]), at least one tracking cell (e.g., tc[p]), at least one sense amplifier (e.g., SA[k]), and a pulse width controller (e.g., 300, FIG. 4 ). The tracking word line comprises a front node (e.g., w1) and a rear node (e.g., w3), and the length between the front node and the rear node is positively correlated to the length of each of the word lines. Each of the memory cells is coupled to one of the at least one bit line and one of the at least one word line. Each of the tracking units is coupled to the tracking bit line. Each of the sense amplifiers is coupled to one of the at least one bit line and receives an enable signal (e.g., GS); each of the sense amplifiers is controlled to be enabled when the enable signal is activated, and is disabled when the enable signal stops being activated. A pulse width controller is coupled to the tracking bit line, the front node, and the rear node, and provides the enable signal. When the voltage of the tracking bit line (e.g., vTBL, FIG. 5) changes to a preset voltage (e.g., vt0, at time t3, FIG. 5), the pulse width controller activates the enable signal (e.g., at time t6, FIG. 5), and changes the voltage of the front node (e.g., vw1, FIG. 5) (e.g., switching from voltage v4 to voltage v3 at time t7, FIG. 5). When the voltage of the front node changes, the tracking word line changes the voltage of the rear node (e.g., vw3, FIG. 5) after a first delay time (e.g., d1, FIG. 5) (e.g., switching from voltage v4 to voltage v3 at time t8, FIG. 5). When the voltage of the rear node changes, the pulse width controller stops activating the enable signal (for example, at time t12, FIG. 5 ) after a second delay time (for example, d2, FIG. 5 ).
一實施例中(例如圖6),該脈寬控制器有一部份(例如邏輯閘G1,圖6)係由一第一供應電壓(例如Vdd1,圖6)供電,各該感測放大器(例如SA[k],圖6)至少有一部份(例如s2[k],圖6)係由一第二供應電壓(例如Vdd2,圖6)供電。一實施例中,該第一供應電壓與該第二供應電壓相異,且該第二延遲時間的長短係負相關於該第二供應電壓的大小。In one embodiment (e.g., FIG. 6 ), a portion of the pulse width controller (e.g., logic gate G1, FIG. 6 ) is powered by a first supply voltage (e.g., Vdd1, FIG. 6 ), and at least a portion of each of the sense amplifiers (e.g., SA[k], FIG. 6 ) (e.g., s2[k], FIG. 6 ) is powered by a second supply voltage (e.g., Vdd2, FIG. 6 ). In one embodiment, the first supply voltage is different from the second supply voltage, and the length of the second delay time is negatively correlated to the magnitude of the second supply voltage.
一實施例中(例如圖6),該脈寬控制器有兩部份(例如邏輯閘G1與延遲電路500,圖6)係分別由兩相異供應電壓(例如Vdd1與Vdd2,圖6)供電,且該第二延遲時間的長短係負相關於該兩相異供應電壓的其中之一(例如Vdd2,圖6)。In one embodiment (e.g., FIG. 6 ), the pulse width controller has two parts (e.g., logic gate G1 and
一實施例中(例如圖4),該脈寬控制器包含一第一邏輯閘(例如G1,圖4)。該第一邏輯閘包含一第一輸入端(例如i1)、一第二輸入端(例如i2)與一第一輸出端(例如o1),分別耦接該追隨位元線、該後節點與一第一節點(例如n1)。該脈寬控制器係於一第三節點(例如n3)形成該致能信號,該第三節點耦接該至少一感測放大器。當該記憶模組進行資料讀取時,該第三節點的電壓係受控於該第一節點的電壓。當該脈寬控制器激發該致能信號時,係使該致能信號由一第一位準(例如v1,圖5)切換至一第二位準(例如v2,圖5);當該脈寬控制器停止激發該致能信號時,係使該致能信號由該第二位準切換回該第一位準。一實施例中(例如圖4),該脈寬控制器更包含一第一反相器(例如iv1,圖4),耦接於該追隨位元線與該第一輸入端之間。一實施例中(例如圖4),該第一邏輯閘係一反及閘。In one embodiment (e.g., FIG. 4 ), the pulse width controller includes a first logic gate (e.g., G1, FIG. 4 ). The first logic gate includes a first input terminal (e.g., i1), a second input terminal (e.g., i2), and a first output terminal (e.g., o1), which are respectively coupled to the tracking bit line, the rear node, and a first node (e.g., n1). The pulse width controller forms the enable signal at a third node (e.g., n3), and the third node is coupled to the at least one sense amplifier. When the memory module performs data reading, the voltage of the third node is controlled by the voltage of the first node. When the pulse width controller activates the enable signal, the enable signal is switched from a first level (e.g., v1, FIG. 5) to a second level (e.g., v2, FIG. 5); when the pulse width controller stops activating the enable signal, the enable signal is switched from the second level back to the first level. In one embodiment (e.g., FIG. 4), the pulse width controller further includes a first inverter (e.g., iv1, FIG. 4), coupled between the tracking bit line and the first input terminal. In one embodiment (e.g., FIG. 4), the first logic gate is a NAND gate.
一實施例中(例如圖4),該脈寬控制器更包含一第三邏輯閘(例如G3,圖4)與一延遲電路(例如500,圖4)。該第三邏輯閘包含一第五輸入端(例如i5)、一第六輸入端(例如i6)與一第三輸出端(例如o3)。該第五輸入端耦接一第二節點(例如n2),該延遲電路耦接於該第二節點與該第六輸入端之間,且該第三輸出端耦接該第三節點。當該記憶模組進行資料讀取時,該第二節點的電壓係受控於該第一節點的電壓。一實施例中(例如圖4),該脈寬控制器更包含一第三反相器(例如iv3,圖4),耦接於該第三輸出端與該第三節點之間。In one embodiment (e.g., FIG. 4 ), the pulse width controller further includes a third logic gate (e.g., G3, FIG. 4 ) and a delay circuit (e.g., 500, FIG. 4 ). The third logic gate includes a fifth input terminal (e.g., i5), a sixth input terminal (e.g., i6), and a third output terminal (e.g., o3). The fifth input terminal is coupled to a second node (e.g., n2), the delay circuit is coupled between the second node and the sixth input terminal, and the third output terminal is coupled to the third node. When the memory module performs data reading, the voltage of the second node is controlled by the voltage of the first node. In one embodiment (eg, FIG. 4 ), the pulse width controller further includes a third inverter (eg, iv3 , FIG. 4 ) coupled between the third output terminal and the third node.
一實施例中(例如圖6),該第一邏輯閘與該延遲電路係分別由一第一供應電壓(例如Vdd1,圖6)與一第二供應電壓供電(例如Vdd2,圖6),且該第一供應電壓與該第二供應電壓相異。一實施例中(例如圖6),各該感測放大器至少有一部分(例如s2[k],圖6)係由該第二供應電壓供電。In one embodiment (e.g., FIG. 6 ), the first logic gate and the delay circuit are powered by a first supply voltage (e.g., Vdd1, FIG. 6 ) and a second supply voltage (e.g., Vdd2, FIG. 6 ), respectively, and the first supply voltage is different from the second supply voltage. In one embodiment (e.g., FIG. 6 ), at least a portion of each sense amplifier (e.g., s2[k], FIG. 6 ) is powered by the second supply voltage.
一實施例中(例如圖4),該脈寬控制器更包含一第二邏輯閘(例如G2,圖4)。該第二邏輯閘包含一第三輸入端(例如i3)、一第四輸入端(例如i4)與一第二輸出端(例如o2),分別耦接該第一節點,一第四節點(例如n4)與該第二節點。In one embodiment (e.g., FIG. 4 ), the pulse width controller further includes a second logic gate (e.g., G2, FIG. 4 ). The second logic gate includes a third input terminal (e.g., i3), a fourth input terminal (e.g., i4), and a second output terminal (e.g., o2), which are respectively coupled to the first node, a fourth node (e.g., n4), and the second node.
一實施例中(例如圖4),該脈寬控制器更包含一第四邏輯閘(例如G4,圖4)。該第四邏輯閘包含一第七輸入端(例如i7)、一第八輸入端(例如i8)與一第四輸出端(例如o4),分別耦接一第一指示信號(例如SCANEN)、一第二指示信號(例如WEI)與該第四節點。一實施例中(例如圖4),該第二邏輯閘、該第三邏輯閘與該第四邏輯閘均係反或閘。In one embodiment (e.g., FIG. 4 ), the pulse width controller further includes a fourth logic gate (e.g., G4, FIG. 4 ). The fourth logic gate includes a seventh input terminal (e.g., i7), an eighth input terminal (e.g., i8), and a fourth output terminal (e.g., o4), which are respectively coupled to a first indication signal (e.g., SCANEN), a second indication signal (e.g., WEI), and the fourth node. In one embodiment (e.g., FIG. 4 ), the second logic gate, the third logic gate, and the fourth logic gate are all negative OR gates.
一實施例中(例如圖4),該記憶模組更包含一有限狀態機電路(例如400,圖4),耦接於該脈寬控制器與該前節點之間。當該脈寬控制器使該前節點的電壓改變時,係使該有限狀態機電路改變該前節點的電壓。In one embodiment (e.g., FIG. 4 ), the memory module further includes a finite state machine circuit (e.g., 400, FIG. 4 ) coupled between the pulse width controller and the previous node. When the pulse width controller changes the voltage of the previous node, the finite state machine circuit changes the voltage of the previous node.
一實施例中(例如圖4),該有限狀態機電路包含一第五節點(例如n5,圖4)與一第六節點(例如n6,圖4);其中,該第五節點耦接該第一節點,且該六節點耦接該前節點。In one embodiment (eg, FIG. 4 ), the finite state machine circuit includes a fifth node (eg, n5, FIG. 4 ) and a sixth node (eg, n6, FIG. 4 ); wherein the fifth node is coupled to the first node, and the sixth node is coupled to the previous node.
一實施例中(例如圖4),該有限狀態機電路更包含一第七節點(例如n7),耦接於一時脈(例如CLK)。一實施例中(例如圖4),該記憶模組更包含一第二反相器(例如iv2,圖4),耦接於該第六節點與該前節點之間。In one embodiment (e.g., FIG. 4 ), the finite state machine circuit further includes a seventh node (e.g., n7) coupled to a clock (e.g., CLK). In one embodiment (e.g., FIG. 4 ), the memory module further includes a second inverter (e.g., iv2, FIG. 4 ) coupled between the sixth node and the previous node.
本發明的目的之一是提供一種可改良感測放大時序適應性的記憶模組(例如100,圖1),其可包含至少一位元線(例如BL[q])、至少一字元線(例如WL[p])、一追隨位元線(例如TBL)、一追隨字元線(例如TWL)、至少一記憶單元(例如c[p,q])、至少一追隨單元(例如tc[p])、至少一感測放大器(例如SA[k])與一第一邏輯閘(例如G1,圖4)。該追隨字元線由一前節點(例如w1)延伸至一後節點(例如w3),且於該前節點至該後節點的長度係正相關於各該字元線的長度。各該記憶單元耦接該至少一位元線的其中之一以及該至少一字元線的其中之一。各該追隨單元耦接該追隨位元線。各該感測放大器耦接該至少一位元線的其中之一,並更耦接於一第三節點(例如n3)。各該感測放大器係在該第三節點的電壓為一第一位準(例如v1,圖5)時失能,並在該第三節點的電壓為一第二位準(例如v2,圖5)時受控致能。該第一邏輯閘包含一第一輸入端(例如i1,圖4)、一第二輸入端(例如i2,圖4)與一第一輸出端(例如o1,圖4),分別耦接該追隨位元線、該後節點與該前節點。當該記憶模組進行資料讀取時,該第三節點的電壓係受控於該第一輸出端的電壓。One of the purposes of the present invention is to provide a memory module (e.g., 100, FIG. 1 ) capable of improving the adaptability of sense amplification timing, which may include at least one bit line (e.g., BL[q]), at least one word line (e.g., WL[p]), a tracking bit line (e.g., TBL), a tracking word line (e.g., TWL), at least one memory cell (e.g., c[p,q]), at least one tracking cell (e.g., tc[p]), at least one sense amplifier (e.g., SA[k]), and a first logic gate (e.g., G1, FIG. 4 ). The tracking word line extends from a front node (e.g., w1) to a rear node (e.g., w3), and the length from the front node to the rear node is positively correlated to the length of each word line. Each memory cell is coupled to one of the at least one bit line and one of the at least one word line. Each of the tracking units is coupled to the tracking bit line. Each of the sensing amplifiers is coupled to one of the at least one bit line and is further coupled to a third node (e.g., n3). Each of the sensing amplifiers is disabled when the voltage of the third node is a first level (e.g., v1, FIG. 5), and is controlled to be enabled when the voltage of the third node is a second level (e.g., v2, FIG. 5). The first logic gate includes a first input terminal (e.g., i1, FIG. 4), a second input terminal (e.g., i2, FIG. 4), and a first output terminal (e.g., o1, FIG. 4), which are respectively coupled to the tracking bit line, the rear node, and the front node. When the memory module performs data reading, the voltage of the third node is controlled by the voltage of the first output terminal.
一實施例中(圖4),該記憶模組更包含一第三邏輯閘(例如G3,圖4)與一延遲電路(例如500,圖4)。該第三邏輯閘包含一第五輸入端(例如i5)、一第六輸入端(例如i6)與一第三輸出端(例如o3);該第五輸入端耦接一第二節點(例如n2),該延遲電路耦接於該第二節點與該第六輸入端之間,該第三輸出端耦接該第三節點,且該第二節點耦接該第一輸出端。一實施例中(例如圖6),該第一邏輯閘與該延遲電路係分別由兩相異供應電壓(例如Vdd1與Vdd2,圖6)供電。In one embodiment (FIG. 4), the memory module further includes a third logic gate (e.g., G3, FIG. 4) and a delay circuit (e.g., 500, FIG. 4). The third logic gate includes a fifth input terminal (e.g., i5), a sixth input terminal (e.g., i6) and a third output terminal (e.g., o3); the fifth input terminal is coupled to a second node (e.g., n2), the delay circuit is coupled between the second node and the sixth input terminal, the third output terminal is coupled to the third node, and the second node is coupled to the first output terminal. In one embodiment (e.g., FIG. 6), the first logic gate and the delay circuit are powered by two phase-differential supply voltages (e.g., Vdd1 and Vdd2, FIG. 6), respectively.
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 better understand the above and other aspects of the present invention, the following embodiments are specifically described in detail with reference to the accompanying drawings:
圖1示意的是依據本發明一實施例的記憶模組100,其可包括P*Q個記憶單元c[1,1]至c[P,Q]、P條字元線WL[1]至WL[P]、Q組位元線BL[1]、BLb[1]至BL[Q]、BLb[Q]、P個追隨單元tc[1]至tc[P]、一輔助字元線w0、一組追隨位元線TBL與TBLb、一追隨字元線TWL、K個緩衝器bf[1]至bf[K]、K個感測放大器SA[1]至SA[K]、K個寫入電路WB[1]至WB[K]、兩週邊電路130與140與一控制電路200。數目P、Q、K可以是預設的整數。數目Q可以是數目K的整數倍;例如,數目Q可以是數目K的一倍、兩倍或四倍等。數目K(感測放大器、寫入電路的個數)代表記憶模組100的輸出入數目。FIG1 shows a
在記憶模組100中,各記憶單元c[p,q](p=1至P,q=1至Q)耦接一對應字元線WL[p]與一組對應位元線BL[q]與BLb[q],並可儲存一位元的資料。延續圖1,圖2示意的是各記憶單元c[p,q]的一實施例;各記憶單元c[p,q]中可包含一組反相器iA[p,q]與iB[p,q],以及一組閘通(pass gate)電晶體A[p,q]與B[p,q](如n通道金氧半電晶體)。反相器iA[p,q]與iB[p,q]形成一閂鎖器L[p,q]。電晶體A[p,q]有一受控端(如閘極端)與兩通道端(如汲極端與源極端),分別耦接字元線WL[p]、位元線BL[q]與閂鎖器L[p,q]的一端。電晶體B[p,q]有一受控端與兩通道端,分別耦接字元線WL[p]、位元線BLb[q]與閂鎖器L[p,q]的另一端。In the
在圖1的記憶模組100中,各追隨單元tc[p]耦接輔助字元線w0以及追隨位元線TBL與TBLb。各追隨單元tc[p]的電路可與各記憶單元c[p,q]的電路相同,使各追隨單元tc[p]可反映(追隨)各記憶單元c[p,q]的性質(如響應時間等)。In the
如圖1所示,在記憶模組100中,追隨字元線TWL可由一節點w1延伸至另一節點w2,再由節點w2延伸至又一節點w3。追隨字元線TWL於節點w1(經節點w2)至節點w3的長度可以正相關於各字元線WL[p]的長度。一實施例中,追隨字元線TWL的線路結構(如線長與線寬)可以趨近或實質等於各字元線WL[p]的線路結構,以反映(追隨)各字元線WL[p]的性質,如等效阻抗、等效負載與響應等。如圖1所示,追隨字元線TWL可由節點w1水平延伸一距離D1至節點w2,再由節點w2反方向延伸相同距離D1(或一近似距離)至節點w3。換言之,追隨字元線TWL於節點w1至w3間的長度可為距離D1的兩倍。如圖1所示,距離D1可以水平跨越位元線BL[J+1]、BLb[J+1]至BL[Q]、BLb[Q]間的空間,而數目J可以是一預設整數;例如,數目Q可以是偶數,數目J則可以等於數目Q的一半。如此,距離D1的兩倍就會等於(或趨近)各字元線WL[p]的長度,因為各字元線WL[p]會水平跨越位元線BL[1]、BLb[1]至BL[J]、BLb[J]以及位元線BL[J+1]、BLb[J+1]至BL[Q]、BLb[Q]。追隨字元線TWL可以不用耦接至任何記憶單元c[p,q];亦即,追隨字元線TWL可以絕緣於記憶單元c[1,1]至c[P,Q]中的任何一個。As shown in FIG1 , in the
週邊電路130耦接字元線WL[1]至WL[P],並受控於控制電路200。週邊電路140耦接各組位元線BL[1]、BLb[1]至BL[Q]、BLb[Q]、感測放大器SA[1]至SA[K]與寫入電路WB[1]至WB[K],亦受控於控制電路200。The
控制電路200接收一時脈CLK,據以控制記憶模組100的運作;控制電路200耦接週邊電路130與140,耦接追隨字元線TWL的節點w1與w3,於一節點n0耦接追隨位元線TBL,並於另一節點n3耦接緩衝器bf[1]至bf[K]。The
控制電路200會由節點n3輸出一信號GS,也就是感測放大器SA[1]至SA[K]的致能信號。各感測放大器SA[k]耦接週邊電路140,也經由對應緩衝器bf[k]耦接節點n3,以接收信號GS。The
當要在一資料寫入回合將資料寫入至一記憶單元c[p,q]時,控制電路200會控制週邊電路130與140,由週邊電路140將對應位元線BL[q]、BLb[q]導通至寫入電路WB[1]至WB[K]的其中之一WB[k],並由週邊電路130驅動對應字元線WL[p],以使記憶單元c[p,q]中的閂鎖器L[p,q](圖2)能被導通至位元線BL[q]與BLb[q]。如此,寫入電路WB[k]便可經由位元線BL[q]與BLb[q]將資料寫入記憶單元c[p,q]。When data is to be written into a memory cell c[p,q] in a data write round, the
延續圖1與圖2,圖3示意的是記憶模組100在進行資料讀取時相關訊號的波形時序。如圖3所示,在時脈CLK的一週期T1中,時脈CLK的電壓會於電壓vck1與vck2間交替一次。當要讀取記憶單元c[p,q](圖1)時,隨著時脈CLK在一時點ta1開始由電壓vck1切換至電壓vck2,控制電路200會控制週邊電路130與140,使週邊電路140將對應位元線BL[q]、BLb[q]導通至感測放大器SA[1]至SA[K]的其中之一SA[k],並在後一時點ta2使週邊電路130將對應字元線WL[p]的電壓vWL[p]由一電壓vWL1(例如一個無法導通閘通電晶體的電壓)驅動至另一電壓vWL2(例如一個足以導通閘通電晶體的電壓),以使記憶單元c[p,q]中的閂鎖器L[p,q](圖2)能被導通至位元線BL[q]與BLb[q]。在另一時點ta3,控制電路200開始激發信號GS,也就是使信號GS開始由一個代表未激發的位準v1切換至另一個代表被激發的位準v2。被激發的信號GS會致能感測放大器SA[1]至SA[K],而被致能的感測器SA[k]便會感測位元線BL[q]與BLb[q]間的電壓差,進而判斷記憶單元c[p,q]中儲存的資料。在後一時點ta4,控制電路200使週邊電路130(圖1)停止驅動字元線WL[p],使電壓vWL[p]開始由電壓vWL2切換回電壓vWL1。在後一時點ta5,控制電路200開始停止激發信號GS,也就是使信號GS開始從位準v2切換回位準v1。隨著信號GS停止激發,感測放大器SA[1]至SA[K]也失能不再運作。在後一時點ta6,時脈CLK的週期T1結束。如圖3所示,信號GS維持於位準v2的脈寬pw1即是感測放大器SA[1]至SA[K]可受控致能的期間。Continuing with FIG. 1 and FIG. 2, FIG. 3 illustrates the waveform timing of the relevant signals when the
延續圖1至圖3,圖4示意的是本發明控制電路200的一實施例。如圖4所示,在本發明的一實施例中,控制電路200可包括一脈寬控制器300、一有限狀態機電路400、兩緩衝器bf1與bf2,以及一反相器iv2。脈寬控制器300中可包括四邏輯閘G1至G4、反相器iv1與iv3以及一延遲電路500。延遲電路500中可包括一緩衝器bf3。有限狀態機電路400可耦接於一供應電壓Vdd1與一地端電壓Vss1之間,並可包含節點n5、n6與n7;時脈CLK可耦接於節點n7。控制電路200尚可包括其他電路元件,但在不影響本發明技術揭露之情形下已省略。Continuing with FIG. 1 to FIG. 3 , FIG. 4 illustrates an embodiment of the
如圖4所示,在控制電路200中,緩衝器bf1耦接於脈寬控制器300的一節點n1與有限狀態機電路400的節點n5之間,反相器iv2與緩衝器bf2串接於有限狀態機電路400的節點n6與追隨字元線TWL的節點w1之間。有限狀態機電路400可在時脈CLK的觸發下控制記憶模組100的運作狀態,脈寬控制器300則可控制信號GS的時序,包括其脈寬pw1(圖3)。As shown in FIG4 , in the
在脈寬控制器300中,邏輯閘G1可以是反及閘,邏輯閘G2至G4可以是反或閘。邏輯閘G1可包括兩輸入端i1、i2與一輸出端o1,邏輯閘G2可包括兩輸入端i3、i4與一輸出端o2,邏輯閘G3可包括兩輸入端i5、i6與一輸出端o3,邏輯閘G4可包括兩輸入端i7、i8與一輸出端o4。反相器iv1耦接於節點n0與輸入端i1之間,延遲電路500耦接於一節點n2與輸入端i6之間,反相器iv3耦接於輸出端o3與節點n3之間,而節點n3的電壓則可形成信號GS。In the
如圖4所示,關於邏輯閘G1,追隨位元線TBL可經由節點n0與反相器iv1耦接輸入端i1,追隨字元線TWL的節點w3可耦接輸入端i2,輸出端o1則可耦接節點n1,而節點n1的電壓可形成一信號TBL_LB。如圖4所示,由節點n1可分枝出兩電路途徑,一電路途徑可使邏輯閘G1的輸出端o1經緩衝器bf1、有限狀態機電路400的節點n5與節點n6、反相器iv2與緩衝器bf2耦接至追隨字元線TWL的節點w1,另一電路途徑則可使邏輯閘G1的輸出端o1經邏輯閘G2、節點n2與延遲電路500、邏輯閘G3與反相器iv3耦接至形成信號GS的節點n3。As shown in FIG4 , regarding the logic gate G1, the following bit line TBL can be coupled to the input terminal i1 via the node n0 and the inverter iv1, the node w3 following the word line TWL can be coupled to the input terminal i2, and the output terminal o1 can be coupled to the node n1, and the voltage of the node n1 can form a signal TBL_LB. As shown in FIG4 , two circuit paths can be branched from the node n1. One circuit path allows the output end o1 of the logic gate G1 to be coupled to the node w1 that follows the word line TWL via the buffer bf1, the node n5 and the node n6 of the finite state machine circuit 400, the inverter iv2 and the buffer bf2. The other circuit path allows the output end o1 of the logic gate G1 to be coupled to the node n3 that forms the signal GS via the logic gate G2, the node n2 and the
在脈寬控制器300中,邏輯閘G2的輸入端i3與輸出端o2可分別耦接節點n1與n2,輸入端i4則可耦接另一節點n4。邏輯閘G3的輸入端i5可耦接節點n2。邏輯閘G4的輸入端i7、i8與輸出端o4可分別耦接一指示信號SCANEN、一指示信號WEI與節點n4。信號SCANEN代表記憶模組100(圖1)是否進行掃描運作。當信號SCANEN為邏輯1時,記憶模組100進行掃描運作;為邏輯0時,記憶模組100不進行掃描運作,故可進行資料讀取或資料寫入。信號WEI則代表記憶模組100是否進行資料讀取。當信號WEI為邏輯0時,記憶模組100進行資料寫入;為邏輯1時,記憶模組100進行資料讀取。In the
圖7a與7b示意的是有限狀態機電路400在一實施例中的狀態列表與狀態圖。如圖7a與7b所示,在本發明的一實施例中,有限狀態機電路400可在狀態S0、P0、Launch、Idle0、Idle1與P1間切換;節點n5與n7(圖4)的邏輯值可視為有限狀態機電路400的輸入,節點n6的邏輯值可視為有限狀態機電路400的輸出,節點n5、n7與n6的邏輯值在圖7a與7b中以xy/z的形式表示。7a and 7b illustrate a state list and a state diagram of the finite state machine circuit 400 in one embodiment. As shown in FIG7a and 7b, in one embodiment of the present invention, the finite state machine circuit 400 can switch between states S0, P0, Launch, Idle0, Idle1 and P1; the logic values of nodes n5 and n7 (FIG4) can be regarded as inputs of the finite state machine circuit 400, and the logic value of node n6 can be regarded as outputs of the finite state machine circuit 400. The logic values of nodes n5, n7 and n6 are represented in the form of xy/z in FIG7a and 7b.
延續圖1至圖4,圖5示意的是當記憶模組100(圖1)在進行資料讀取時控制電路200(圖4)中各相關信號的波形時序,其中,電壓vw1、vw3、vTBL、vi5與vi6分別為節點w1、節點w3、追隨位元線TBL(節點n0)、輸入端i5與i6的電壓。在一時點t0,時脈CLK開始由電壓vck1(例如一代表邏輯0的電壓)切換至電壓vck2(例如一代表邏輯1的電壓)。當時脈CLK由電壓vck1切換至電壓vck2時,有限狀態機電路400(圖4)可使節點n6的電壓由供應電壓Vdd1(其可代表邏輯1)改變至地端電壓Vss1(其可代表邏輯0)。隨著節點n6的電壓改變,反相器iv2與緩衝器bf2的運作會在一稍後時點t1使節點w1的電壓vw1(圖5)開始由電壓v3(例如一代表邏輯0的電壓)切換至電壓v4(例如一代表邏輯1的電壓)。由於追隨字元線TWL的長度,節點w3的電壓vw3會在一段延遲時間d1後由電壓v3切換至電壓v4。Continuing with FIG. 1 to FIG. 4 , FIG. 5 illustrates the waveform timing of each relevant signal in the control circuit 200 ( FIG. 4 ) when the memory module 100 ( FIG. 1 ) is reading data, wherein voltages vw1, vw3, vTBL, vi5, and vi6 are voltages of node w1, node w3, tracking bit line TBL (node n0), and input terminals i5 and i6, respectively. At a time point t0, the clock CLK starts to switch from voltage vck1 (e.g., a voltage representing logic 0) to voltage vck2 (e.g., a voltage representing logic 1). When the clock CLK switches from voltage vck1 to voltage vck2, the finite state machine circuit 400 (FIG. 4) can change the voltage of node n6 from supply voltage Vdd1 (which can represent logic 1) to ground voltage Vss1 (which can represent logic 0). As the voltage of node n6 changes, the operation of inverter iv2 and buffer bf2 causes the voltage vw1 of node w1 (FIG. 5) to start switching from voltage v3 (e.g., a voltage representing logic 0) to voltage v4 (e.g., a voltage representing logic 1) at a later time t1. Due to the length of the tracking word line TWL, the voltage vw3 of the node w3 switches from voltage v3 to voltage v4 after a delay time d1.
隨著追隨字元線TWL的電壓切換至電壓v4,追隨位元線TBL於節點n0的電壓vTBL會在一時點t2後由一初始電壓vpr0開始改變(例如下降),並在另一時點t3改變至一預設電壓vt0。舉例而言,電壓vt0可以是反相器iv1的翻轉電壓;當節點n0的電壓還未改變至此電壓vt0前(即時點t3前),反相器iv1會將節點n0的電壓判定為邏輯1;當節點n0的電壓改變至此電壓vt0後(即時點t3後),反相器iv1便轉而將節點n0的電壓判定為邏輯0。因此,反相器iv1會在時點t3後將輸出至輸入端i1的邏輯0切換為邏輯1。響應輸入端i1的輸入改變,邏輯閘G1會在後一時點t4使節點n1的信號TBL_LB開始由電壓v6(例如說是一代表邏輯1的電壓)切換至電壓v5(例如說是一代表邏輯0的電壓)。As the voltage of the tracking word line TWL switches to the voltage v4, the voltage vTBL of the tracking bit line TBL at the node n0 will start to change (e.g., decrease) from an initial voltage vpr0 after a time point t2, and change to a preset voltage vt0 at another time point t3. For example, the voltage vt0 can be the flip voltage of the inverter iv1; before the voltage of the node n0 has not changed to this voltage vt0 (i.e., before the time point t3), the inverter iv1 will determine the voltage of the node n0 as a logical 1; after the voltage of the node n0 changes to this voltage vt0 (i.e., after the time point t3), the inverter iv1 will switch to determine the voltage of the node n0 as a logical 0. Therefore, the inverter iv1 switches the
在進行資料讀取期間,信號SCANEN(圖4)與WEI會使邏輯閘G4持續由輸出端o4輸出邏輯0至邏輯閘G2的輸入端i4。因此,當信號TBL_LB在時點t4開始由電壓v6切換至電壓v5時,邏輯閘G2會在後一時點t5使輸入端i5的電壓vi5(圖5)開始由電壓v7(例如一代表邏輯0的電壓)切換至電壓v8(例如一代表邏輯1的電壓)。當輸入端i5的電壓vi5(圖5)在時點t5開始由電壓v7切換為電壓v8時,由於節點n2與輸入端i6間有延遲電路500(圖4),輸入端i6的電壓vi6仍會維持於電壓v7。因此,隨著電壓vi5在時點t5開始由電壓v7切換至電壓v8,邏輯閘G3與反相器iv3的運作會使節點n3的信號GS在一稍後時點t6開始由位準v1(例如一代表邏輯0的位準)切換至位準v2(例如一代表邏輯1的位準),也就是開始激發信號GS。亦即,隨著追隨位元線TBL在時點t3改變至預設電壓vt0,反相器iv1、邏輯閘G1與邏輯閘G2、G3與反相器iv3會被帶動進行連串運作,進而在時點t6開始激發信號GS。During data reading, the signals SCANEN (FIG. 4) and WEI will cause the logic gate G4 to continuously
當節點n1的信號TBL_LB在時點t4開始由電壓v6切換至電壓v5時,緩衝器bf1會使節點n5的電壓也隨之切換;響應節點n5的電壓切換,有限狀態機電路400會使節點n6的電壓改變至供應電壓Vdd1;反相器iv2與緩衝器bf2的運作會在一時點t7使節點w1的電壓vw1開始由電壓v4切換回電壓v3。亦即,隨著追隨位元線TBL在時點t3改變至預設電壓vt0,反相器iv1、邏輯閘G1與緩衝器bf1、有限狀態機電路400、反相器iv2與緩衝器bf2也會被帶動進行連串運作,進而在時點t7使節點w1的電壓vw1開始改變(由電壓v4切換回電壓v3)。When the signal TBL_LB of the node n1 starts to switch from the voltage v6 to the voltage v5 at the time t4, the buffer bf1 will cause the voltage of the node n5 to switch accordingly; in response to the voltage switching of the node n5, the finite state machine circuit 400 will change the voltage of the node n6 to the supply voltage Vdd1; the operation of the inverter iv2 and the buffer bf2 will cause the voltage vw1 of the node w1 to start switching from the voltage v4 back to the voltage v3 at a time t7. That is, as the tracking bit line TBL changes to the preset voltage vt0 at time t3, the inverter iv1, the logic gate G1 and the buffer bf1, the finite state machine circuit 400, the inverter iv2 and the buffer bf2 are also driven to operate in series, thereby causing the voltage vw1 of the node w1 to start changing at time t7 (switching from voltage v4 back to voltage v3).
當節點w1的電壓vw1在時點t7開始改變,追隨字元線TWL會在延遲時間d1後的另一時點t8(=t7+d1)使節點w3的電壓vw3也開始改變,由電壓v4切換回電壓v3。隨著追隨字元線TWL的電壓回到電壓v3,追隨位元線TBL的電壓vTBL可被回復至電壓vpr0。When the voltage vw1 of node w1 starts to change at time t7, the tracking word line TWL will cause the voltage vw3 of node w3 to also start to change at another time t8 (= t7 + d1) after a delay time d1, switching from voltage v4 back to voltage v3. As the voltage of the tracking word line TWL returns to voltage v3, the voltage vTBL of the tracking bit line TBL can be restored to voltage vpr0.
隨著電壓vw3在時點t8開始由電壓v4切換回電壓v3,邏輯閘G1會在一稍後時點t9使信號TBL_LB開始由電壓v5切換回電壓v6。連帶地,邏輯閘G2會在後一時點t10使輸入端i5的電壓vi5開始由電壓v8切換回電壓v7,延遲電路500則在延遲時間d2後的另一時點t11(=t10+d2)使輸入端i6的電壓vi6開始由電壓v8切換回電壓v7。隨著輸入端i5與i6的電壓vi5與vi6都切換回電壓v7,邏輯閘G3與反相器iv3的運作會在一時點t12使信號GS開始由位準v2切換回位準v1,停止激發信號GS。亦即,隨著節點w3的電壓vw3在時點t8開始改變,邏輯閘G1、G2、延遲電路500、邏輯閘G3與反相器iv3會被帶動進行連串運作,在時點t12開始停止激發信號GS。As the voltage vw3 starts to switch from the voltage v4 back to the voltage v3 at the time t8, the logic gate G1 will make the signal TBL_LB start to switch from the voltage v5 back to the voltage v6 at a later time t9. In conjunction, the logic gate G2 will make the voltage vi5 of the input terminal i5 start to switch from the voltage v8 back to the voltage v7 at the later time t10, and the
由以上描述可知,本發明對信號GS的脈寬控制可簡述如下。當追隨位元線TBL的電壓vTBL改變至預設電壓vt0時(時點t3),脈寬控制器300會在時點t6激發致能信號GS(藉由反相器iv1、邏輯閘G1、G2、G3與反相器iv3在時點t3至t6的運作),並在時點t7使節點w1的電壓vw1改變(藉由反相器iv1、邏輯閘G1、緩衝器bf1、有限狀態機電路400、反相器iv2與緩衝器bf2在時點t3至t7的運作)。當節點w1的電壓vw1在時點t7改變,追隨字元線TWL在延遲時間d1後的時點t8使節點w3的電壓vw3改變。當節點w3的電壓vw3在時點t8改變,脈寬控制器300在延遲時間d2後的時點t12停止激發致能信號GS(藉由邏輯閘G1、G2、延遲電路500、邏輯閘G3與反相器iv3在時點t8至t12的運作)。From the above description, the pulse width control of the signal GS of the present invention can be briefly described as follows. When the voltage vTBL of the tracking bit line TBL changes to the preset voltage vt0 (time point t3), the
由圖5可知,經由脈寬控制器300的運作,從「開始激發信號GS」(時點t6)到「開始停止激發信號GS」(時點t12)間的時距(以下稱為激發時距AT1)會涵蓋延遲時間d1與d2之和。由於「開始激發信號GS」後還需要一段暫態時間(例如上升時間)才能使信號GS由未激發位準(例如邏輯0的位準)真正到達已激發位準(例如邏輯1的位準),信號GS真正維持於受激發位準的脈寬係取決於激發時距減去暫態時間的結果。由於信號GS需被傳送至所有感測放大器SA[1]至SA[K](經由緩衝器bf[1]至bf[K]),若記憶模組的輸出入數目(感測放大器的數目K)越多,激發信號GS所面臨的阻抗越大,連帶地,暫態時間也會較長。As shown in FIG5 , through the operation of the
在習知技術中,「開始激發信號GS」到「開始停止激發信號GS」間的激發時距是固定的;因此,若記憶模組的輸出入數目較大,暫態時間較長,信號GS的脈寬(固定激發時距與較長暫態時間之差)就會相對變短,導致信號GS脈寬不足的問題。In conventional technology, the excitation time interval between "start excitation signal GS" and "start stop excitation signal GS" is fixed; therefore, if the number of inputs and outputs of the memory module is large and the transient time is long, the pulse width of the signal GS (the difference between the fixed excitation time interval and the longer transient time) will become relatively shorter, resulting in the problem of insufficient pulse width of the signal GS.
相較於先前技術的固定激發時距,在本發明技術中,「開始激發信號GS」到「開始停止激發信號GS」間的激發時距AT1則是自適應的。若記憶模組的輸出入數目K較大,各字元線WL[p]與追隨字元線TWL(圖1)的長度會較長,故延遲時間d1(圖5)也會較長,涵蓋延遲時間d1的激發時距AT1也就隨之變長。如此一來,即使暫態時間較長,信號GS的脈寬(較長激發時距與較長暫態時間之差)也不會不足。Compared to the fixed excitation time interval of the prior art, in the present invention, the excitation time interval AT1 between "start excitation signal GS" and "start stop excitation signal GS" is adaptive. If the number of inputs and outputs K of the memory module is large, the length of each word line WL[p] and the tracking word line TWL (Figure 1) will be longer, so the delay time d1 (Figure 5) will also be longer, and the excitation time interval AT1 covering the delay time d1 will also become longer. In this way, even if the transient time is longer, the pulse width of the signal GS (the difference between the longer excitation time interval and the longer transient time) will not be insufficient.
在記憶模組100中,控制電路200本來就會運用追隨字元線TWL追隨各字元線WL[p]的特性與表現,據以動態調整對記憶模組100的控制。本發明脈寬控制器300則進一步擴展追隨字元線TWL的用途,利用既有的追隨字元線TWL進行回授(由輸出端o1至輸入端i2),以使激發時距AT1長短得以正相關於輸出入數目K,進而確保信號GS具有足夠的脈寬。如此,本發明脈寬控制器300便能一體適用於不同輸出入數目的各種記憶模組。再者,脈寬控制器300中的延遲電路500也不需要太多的串接反相器來延長激發時距AT1,進而減少脈寬控制器300的佈局面積。一實施例中,延遲電路500中的緩衝器bf3可以簡單由兩個串接反相器(未繪示)形成。In the
延續圖1至圖5,圖6示意的是依據本發明一實施例的供應電壓配置。因應功耗、效能及/或運作上的特殊需求,某些記憶模組會跨越不同電源領域(power domain),不同的部份會分別使用不同供應電壓。例如,如圖6所示,各感測放大器SA[k](k=1至K)可以包括兩部份s1[k]與s2[k],分別屬於兩個電源領域;部份s1[k]可偏壓於供應電壓Vdd1與地端電壓Vss1之間,由供應電壓Vdd1(例如一記憶體供應電壓)供電,部份s2[k]則可偏壓於另一供應電壓Vdd2與另一地端電壓Vss2之間,由供應電壓Vdd2(例如一介面供應電壓)供電。Continuing with Figures 1 to 5, Figure 6 illustrates a supply voltage configuration according to an embodiment of the present invention. In response to special requirements on power consumption, performance and/or operation, some memory modules will span different power domains, and different parts will use different supply voltages. For example, as shown in FIG. 6 , each sense amplifier SA[k] (k=1 to K) may include two parts s1[k] and s2[k], which belong to two power domains respectively; the part s1[k] may be biased between a supply voltage Vdd1 and a ground voltage Vss1, and powered by the supply voltage Vdd1 (e.g., a memory supply voltage), and the part s2[k] may be biased between another supply voltage Vdd2 and another ground voltage Vss2, and powered by the supply voltage Vdd2 (e.g., an interface supply voltage).
在一採用雙軌(dual-rail)感測放大機制的實施例中,供應電壓Vdd1與Vdd2是相異的,供應電壓Vdd1可以大於或小於供應電壓Vdd2,地端電壓Vss1與Vss2則可以是耦接在一起的。由於各感測放大器SA[k]的部份s2[k]是由供應電壓Vdd2供電的,感測放大器SA[k]的運作速度會與供應電壓Vdd2的高低相關;連帶地,信號GS的脈寬需求也會與供應電壓Vdd2的高低相關。若供應電壓Vdd2較低,感測放大器SA[k]的運作速度較慢,信號GS的脈寬也應該要變長,讓較慢的感測放大器SA[k]有較長的時間進行感測;若供應電壓Vdd2較高,感測放大器SA[k]的運作速度較快,信號GS的脈寬也可隨之縮短。In an embodiment using a dual-rail sense amplifier mechanism, supply voltages Vdd1 and Vdd2 are different, supply voltage Vdd1 can be greater than or less than supply voltage Vdd2, and ground voltages Vss1 and Vss2 can be coupled together. Since portion s2[k] of each sense amplifier SA[k] is powered by supply voltage Vdd2, the operating speed of the sense amplifier SA[k] is related to the level of supply voltage Vdd2; and, as a result, the pulse width requirement of signal GS is also related to the level of supply voltage Vdd2. If the supply voltage Vdd2 is lower, the sense amplifier SA[k] operates slower and the pulse width of the signal GS should be longer so that the slower sense amplifier SA[k] has more time to sense. If the supply voltage Vdd2 is higher, the sense amplifier SA[k] operates faster and the pulse width of the signal GS can be shortened accordingly.
為了適應感測放大器SA[k]的供應電壓Vdd2,脈寬控制器300內的延遲電路500(緩衝器bf3)可以和感測放大器SA[k]的部份s2[k]一樣偏壓於供應電壓Vdd2與地端電壓Vss2之間,由供應電壓Vdd2供電。另一方面,脈寬控制器300中內的邏輯閘G1至G4與反相器iv1與iv3,乃至於控制電脈路200中的有限狀態機電路400、反相器iv2與緩衝器bf1、bf2則可以和感測放大器SA[k]的部份s1[k]一樣偏壓於供應電壓Vdd1與地端電壓Vss1之間,由供應電壓Vdd1供電。如此,延遲電路500的延遲時間d2就會負相關於供應電壓Vdd2的高低;若供應電壓Vdd2較低,延遲電路500的運作速度較慢,延遲時間d2也隨之延長;若供應電壓Vdd2較高,延遲電路500的運作速度較快,延遲時間d2也隨之縮短。由於信號GS的激發時距AT1涵蓋了延遲電路500的延遲時間d2(圖5),激發時距AT1長短也會和延遲時間d2一樣負相關於供應電壓Vdd2的高低;若供應電壓Vdd2較低,激發時距AT1與信號GS的脈寬會適應性地延長;若供應電壓Vdd2較高,激發時距AT1與信號GS的脈寬則會適應性地縮短。To adapt to the supply voltage Vdd2 of the sense amplifier SA[k], the delay circuit 500 (buffer bf3) in the
在雙軌感測放大機制的另一實施例中,緩衝器bf1及/或bf2也可以偏壓於供應電壓Vdd2與地端電壓Vss2之間,與輔助延遲電路500一起使激發時距AT1(圖5)負相關於供應電壓Vdd2。又一實施例中,緩衝器bf1及/或bf2偏壓於供應電壓Vdd2與地端電壓Vss2之間,延遲電路500可省略,邏輯閘G3可以是一反相器,耦接於節點n2與反相器iv3之間。In another embodiment of the dual-rail sense amplifier mechanism, the buffer bf1 and/or bf2 may also be biased between the supply voltage Vdd2 and the ground voltage Vss2, and together with the
相較於雙軌感測放大機制,在另一類的單軌感測放大機制的實施例中,供應電壓Vdd2可以等於供應電壓Vdd1,亦即,控制電路200與各感測放大器SA[k]均屬於同一電源領域,統一由供應電壓Vdd1供電。Compared to the dual-rail sense amplifier mechanism, in another type of single-rail sense amplifier mechanism embodiment, the supply voltage Vdd2 can be equal to the supply voltage Vdd1, that is, the
總結來說,在習知技術中,感測放大器的致能信號受限於固定的激發時距,無法適應各種記憶模組的不同需求。相較之下,本發明技術可擴展既有追隨字元線的用途,在本發明脈寬控制器中利用追隨字元線的回授使激發時距可以正相關於輸出入數目,以適應不同輸出入數目的各種記憶模組。再者,因應雙軌感測放大機制,本發明脈寬控制器的不同部份也可以分別屬於不同電源領域,使致能信號的激發時距可以負相關於感測放大器的供應電壓,進而適應不同供應電壓配置的各種記憶模組。In summary, in the prior art, the enable signal of the sense amplifier is limited by a fixed excitation time interval and cannot adapt to the different needs of various memory modules. In contrast, the technology of the present invention can expand the use of existing tracking word lines. In the pulse width controller of the present invention, the feedback of the tracking word line is used to make the excitation time interval positively related to the number of inputs and outputs, so as to adapt to various memory modules with different numbers of inputs and outputs. Furthermore, in response to the dual-track sense amplifier mechanism, different parts of the pulse width controller of the present invention can also belong to different power supply domains, so that the excitation time interval of the enable signal can be negatively related to the supply voltage of the sense amplifier, thereby adapting to various memory modules with different supply voltage configurations.
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。In summary, although the present invention has been disclosed as above by way of embodiments, it is not intended to limit the present invention. A person having ordinary knowledge in the technical field to which the present invention belongs may 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 shall be subject to the scope defined in the attached patent application.
100:記憶模組 130、140:週邊電路 200:控制電路 300:脈寬控制器 400:有限狀態機電路 500:延遲電路 WL[1]-WL[P]:字元線 BL[1]-BL[Q]、BLb[1]-BLb[Q]:位元線 c[1,1]-c[P,Q]:記憶單元 TWL:追隨字元線 CLK:時脈 D1:距離 TBL、TBLb:追隨位元線 w0:輔助字元線 tc[1]-tc[P]:追隨單元 bf[1]-bf[K]、bf1-bf3:緩衝器 SA[1]-SA[K]:感測放大器 WB[1]-WB[K]:寫入電路 A[p,q]、B[p,q]:電晶體 n0-n7、w1-w3:節點 T1:週期 pw1:脈寬 G1-G4:邏輯閘 i1-i8:輸入端 o1-o4:輸出端 iv1-iv3、iA[p,q]、iB[p,q]:反相器 L[p,q]:閂鎖器 GS、TBL_LB、SCANEN、WEI:信號 v1-v2:位準 vck1-vck2、vWL1-vWL2、v3-v8、vpr0、vt0:電壓 vWL[p]、vw1、vw3、vTBL、vi5、vi6:電壓 AT1:激發時距 ta1-ta6、t0-t12:時點 Vdd1、Vdd2:供應電壓 Vss1、Vss2:地端電壓 s1[1]-s1[K]、s2[1]-s2[K]:部份 S0、P0、Launch、Idle0、P1、Idle1:狀態 x、y、z:邏輯值 100: memory module 130, 140: peripheral circuit 200: control circuit 300: pulse width controller 400: finite state machine circuit 500: delay circuit WL[1]-WL[P]: word line BL[1]-BL[Q], BLb[1]-BLb[Q]: bit line c[1,1]-c[P,Q]: memory cell TWL: tracking word line CLK: clock D1: distance TBL, TBLb: tracking bit line w0: auxiliary word line tc[1]-tc[P]: tracking cell bf[1]-bf[K], bf1-bf3: buffer SA[1]-SA[K]: sense amplifier WB[1]-WB[K]: write circuit A[p,q], B[p,q]: transistor n0-n7, w1-w3: node T1: cycle pw1: pulse width G1-G4: logic gate i1-i8: input o1-o4: output iv1-iv3, iA[p,q], iB[p,q]: inverter L[p,q]: latch GS, TBL_LB, SCANEN, WEI: signal v1-v2: level vck1-vck2, vWL1-vWL2, v3-v8, vpr0, vt0: voltage vWL[p], vw1, vw3, vTBL, vi5, vi6: voltage AT1: triggering time ta1-ta6, t0-t12: timing Vdd1, Vdd2: supply voltage Vss1, Vss2: ground voltage s1[1]-s1[K], s2[1]-s2[K]: part S0, P0, Launch, Idle0, P1, Idle1: status x, y, z: logical value
圖1示意的是依據本發明一實施例的記憶模組,其可包含複數個記憶單元與一控制電路。 圖2示意的是圖1記憶單元的一實施例。 圖3示意的是圖1記憶模組在進行資料讀取時相關波形時序的一實施例。 圖4示意的是圖1控制電路的一實施例。 圖5示意的是圖4控制電路中相關波形時序的一實施例。 圖6示意的是圖4中供應電壓配置的一實施例。 圖7a與7b示意的是圖4中有限狀態機電路於一實施例中的狀態列表與狀態圖。 FIG. 1 illustrates a memory module according to an embodiment of the present invention, which may include a plurality of memory cells and a control circuit. FIG. 2 illustrates an embodiment of the memory cell of FIG. 1. FIG. 3 illustrates an embodiment of the waveform timing of the memory module of FIG. 1 when reading data. FIG. 4 illustrates an embodiment of the control circuit of FIG. 1. FIG. 5 illustrates an embodiment of the waveform timing of the control circuit of FIG. 4. FIG. 6 illustrates an embodiment of the supply voltage configuration in FIG. 4. FIG. 7a and FIG. 7b illustrate a state list and a state diagram of the finite state machine circuit in FIG. 4 in an embodiment.
200:控制電路 300:脈寬控制器 400:有限狀態機電路 500:延遲電路 TBL:追隨位元線 TWL:追隨字元線 n0-n7、w1-w3:節點 G1-G4:邏輯閘 i1-i8:輸入端 o1-o4:輸出端 iv1-iv3:反相器 bf[1]-bf[K]、bf1-bf3:緩衝器 SA[1]-SA[K]:感測放大器 CLK:時脈 GS、TBL_LB、SCANEN、WEI:信號 Vdd1:供應電壓 Vss1:地端電壓 200: Control circuit 300: Pulse width controller 400: Finite state machine circuit 500: Delay circuit TBL: Follow bit line TWL: Follow word line n0-n7, w1-w3: Node G1-G4: Logic gate i1-i8: Input o1-o4: Output iv1-iv3: Inverter bf[1]-bf[K], bf1-bf3: Buffer SA[1]-SA[K]: Sense amplifier CLK: Clock GS, TBL_LB, SCANEN, WEI: Signal Vdd1: Supply voltage Vss1: Ground voltage
Claims (20)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW113132059A TWI874275B (en) | 2022-04-13 | 2022-04-13 | Memory module with improved timing adaptivity of sensing amplification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW113132059A TWI874275B (en) | 2022-04-13 | 2022-04-13 | Memory module with improved timing adaptivity of sensing amplification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TW202501474A TW202501474A (en) | 2025-01-01 |
| TWI874275B true TWI874275B (en) | 2025-02-21 |
Family
ID=95152720
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW113132059A TWI874275B (en) | 2022-04-13 | 2022-04-13 | Memory module with improved timing adaptivity of sensing amplification |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TWI874275B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW202022859A (en) * | 2018-12-14 | 2020-06-16 | 日商東芝記憶體股份有限公司 | Semiconductor memory device |
| TW202109535A (en) * | 2019-08-28 | 2021-03-01 | 日商鎧俠股份有限公司 | Nonvolatile semiconductor memory device |
| US20220020420A1 (en) * | 2020-07-15 | 2022-01-20 | Synopsys, Inc. | Sequential delay enabler timer circuit for low voltage operation for srams |
| US20220068354A1 (en) * | 2020-08-26 | 2022-03-03 | Micron Technology, Inc. | Sense amplifier operation at edge memory array mats |
-
2022
- 2022-04-13 TW TW113132059A patent/TWI874275B/en active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW202022859A (en) * | 2018-12-14 | 2020-06-16 | 日商東芝記憶體股份有限公司 | Semiconductor memory device |
| TW202109535A (en) * | 2019-08-28 | 2021-03-01 | 日商鎧俠股份有限公司 | Nonvolatile semiconductor memory device |
| US20220020420A1 (en) * | 2020-07-15 | 2022-01-20 | Synopsys, Inc. | Sequential delay enabler timer circuit for low voltage operation for srams |
| US20220068354A1 (en) * | 2020-08-26 | 2022-03-03 | Micron Technology, Inc. | Sense amplifier operation at edge memory array mats |
Also Published As
| Publication number | Publication date |
|---|---|
| TW202501474A (en) | 2025-01-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR100801059B1 (en) | Driver Circuit of Semiconductor Memory Device to Reduce Leakage Current | |
| JP6353668B2 (en) | Semiconductor memory device | |
| JPH0650599B2 (en) | Semiconductor memory | |
| KR20010073605A (en) | High voltage discharge circuit of a semiconductor memory device | |
| US20220399052A1 (en) | Circuit module with reliable margin configuration | |
| JP4472839B2 (en) | Flash memory device | |
| JP2586723B2 (en) | Sense amplifier | |
| KR980011453A (en) | Output buffer circuit | |
| JP3003631B2 (en) | Nonvolatile semiconductor memory device | |
| KR20000020260A (en) | Synchronous semiconductor memory device | |
| US20050127945A1 (en) | Data inversion circuits having a bypass mode of operation and methods of operating the same | |
| US5841702A (en) | Output circuit for memory device | |
| JP3957823B2 (en) | Voltage detection circuit | |
| JPH0745075A (en) | Semiconductor integrated circuit | |
| JP5763659B2 (en) | Semiconductor memory device | |
| TWI874275B (en) | Memory module with improved timing adaptivity of sensing amplification | |
| KR100304709B1 (en) | Semiconductor Memory Device Whose Data Input/Output Mode can be Controled outside | |
| TWI858922B (en) | Method applied to memory module for improving timing adaptivity of sensing amplification | |
| KR100492907B1 (en) | Memory device with modified global input output scheme | |
| KR100343138B1 (en) | Semiconductor Memory Device Having Write Masking Function and the Write Masking Method | |
| JPH10290152A (en) | Data output buffer circuit | |
| TWI819567B (en) | Memory module with improved timing adaptivity of sensing amplification | |
| US6222787B1 (en) | Integrated circuit memory devices having improved sense and restore operation reliability | |
| JP6578413B2 (en) | Semiconductor device | |
| JPH08340245A (en) | Signal output circuit and semiconductor integrated circuit |