Low power RAM memory cell using a precharge line pulse during write operation

[0002] This application is based upon and claims priority from prior European Patent Application No. 97-120943.2, filed Nov. 28, 1997, the entire disclosure of which is herein incorporated by reference.

[0003] 1. Field of the Invention

[0004] The present invention relates to a semiconductor memory device, and more specifically to a low power RAM memory cell having cross-coupled CMOS inverters.

[0005] 2. Description of the Related Art

[0006] A conventional semiconductor static random access memory (SRAM) device is formed with static memory cells that each have six transistors. FIG. 1 shows a conventional CMOS six transistor SRAM memory cell. The memory cell 1 includes a pair of cross-coupled CMOS inverters 2 and 3, each of which is coupled to a bit line 4 and 5. In particular, the first inverter 2 is coupled to a first bit line 4 through a bi-directional access device 6, and the second inverter 3 is coupled to an adjacent second bit line 5 through a second access device 7. During reading and writing operations, different voltages must be applied to the two bit lines 4 and 5. Thus, this type of access to the storage node of the memory cell can be termed “differential.”

[0007] More specifically, during reading from the memory cell of FIG. 1, the bit line voltage swing amplitude is dependent upon the length of time the memory cell has been activated. The voltage difference caused by the swing can be kept quite small and sensed by the sense amplifier of the memory device in order to reduce power consumption. Further, during writing to the memory cell, the bit line voltage swing is made as large as possible (e.g., the full CMOS logic voltage level) in order to toggle (i.e., write to) the memory cell. Thus, in an SRAM six transistor memory cell array with m rows and n columns, the current consumption during reading and writing can be estimated using the following formulas:

[0008] Idd_r=n*m*Cb*ΔV/dt  (1)

[0009] Idd_w=n*m*Cb*ΔV_w/dt  (2)

[0010] where n is the number of bits in the word being read or written, Cb is the bit line capacitance associated with a given cell, ΔV_ris the bit line voltage swing during a read operation, and ΔV_wis the bit line voltage swing during a write operation. Typically, ΔV_wcorresponds to the supply voltage level Vdd.

[0011] Previous efforts to reduce the power consumed by such a memory matrix focus on changing one or more of the parameters in the above formulas. One such technique is disclosed by N. Kushiyama et al. in “A 295 MHz CMOS 1M (×256) embedded SRAM using I-directional read/write shared sense amplifiers and self-timed pulsed word-line drivers” (ISSCC Dig. Tech. Papers, February 1995, pages 182-183). According to this technique, power consumption is reduced by reducing the number of cells on the bit line through a hierarchical bit line scheme.

[0012] Another power reduction technique is disclosed by B. Amrutur and H. Horowitz in “Technique to reduce power in fast wide memories” (Dig. Tech. Papers, October 1994, Symp. on Low Power Electronics, pages 92-93). This technique reduces power consumption by limiting the bit line voltage swing during a read by controlling the word line pulse length. Yet another power reduction technique is disclosed by T. Blalock and R. Jager in “A high-speed clamped bit line current-mode sense amplifier” (IEEE J. Solid State Circuits, Vol. 26, No. 4, April 1991, pages 542-548). This solution also reduces power consumption by limiting the bit line voltage swing during a read, but does so using current-mode sense amplifiers so as to reduce ΔV_r. Still another power reduction technique limits the bit line voltage swing during a write to a predetermined value (i.e., Vdd−Vt) using NMOS transistors during precharging.

[0013] In view of these drawbacks, it is an object of the present invention to remove the above-mentioned drawbacks and to provide a RAM memory device with reduced power consumption. A RAM memory device is formed with memory cells that each have cross-coupled inverters, and the pull-down transistors of the inverters are connected to a precharge line. During operation, the source terminals of the pull-down transistors in the memory cell are precharged to “reset” the internal nodes of the memory cell to a predetermined state. Then, data can be stored by creating a small voltage difference between bit lines and coupling the bit lines to the internal nodes of the memory cell. In an alternative embodiment of the present invention, the pull-up transistors of the inverters are connected to a precharge line so that the source terminals of the pull-up transistors in the memory cell are precharged.

[0014] Another object of the present invention is to provide a RAM memory device that has relatively long word length but reduced power consumption.

[0015] A further object of the present invention is to provide a technique for reducing voltage swing on the memory array bit lines during a write operation.

[0016] Yet another object of the present invention is to provide a RAM memory device having a simple write operation.

[0017] One embodiment of the present invention provides a semiconductor memory cell that includes a word line, two bit lines, a precharge line, and two cross-coupled inverters. Each inverter is formed by a P-channel transistor and an N-channel transistor. Additionally, a first access transistor selectively couples one bit line to the output of one inverter, and a second access transistor selectively couples the other bit line to the output of the other inverter. One terminal of the N-channel transistor of each of the inverters is connected to the precharge line. In a preferred embodiment, the access transistors are P-channel transistors and the gate terminal of each PMOS access transistor is connected to the word line. Additionally, the present invention provides a method of writing data to a semiconductor memory cell that is connected to a pair of bit lines.

[0018] Other objects, features, and advantages of the present invention will become apparent from the following detailed description. It should be understood, however, that the detailed description and specific examples, while indicating preferred embodiments of the present invention, are given by way of illustration only and various modifications may naturally be performed without deviating from the present invention.

[0019] FIG. 1 is a schematic diagram of a conventional CMOS SRAM memory cell;

[0020] FIG. 2 is a schematic diagram of a memory cell according to a first embodiment of the present invention;

[0021] FIG. 3 is a schematic diagram of a portion of a memory array that includes the memory cells of FIG. 2;

[0022] FIG. 4 is a timing diagram for a write operation in the memory cell of FIG. 2;

[0023] FIG. 5 is a schematic diagram of a memory cell according to a second embodiment of the present invention;

[0024] FIG. 6 is a schematic diagram of a memory cell according to a third embodiment of the present invention;

[0025] FIG. 7 is a schematic diagram of a memory cell according to a fourth embodiment of the present invention; and

[0026] FIG. 8 is a schematic diagram of a dual port memory cell according to a fifth embodiment of the present invention.

[0027] Preferred embodiments of the present invention will be described in detail hereinbelow with reference to the attached drawings.

[0028] FIG. 2 shows a static random access memory (SRAM) cell according to a preferred embodiment of the present invention. The memory cell 10, which is formed using CMOS technology, is designed for a semiconductor SRAM memory device of the type having six transistor memory cells. In particular, the memory cell 10 of FIG. 2 is integrated into a memory array or matrix that includes m rows and n columns, as partially shown in FIG. 3. Rows are identified with word lines WL and columns are identified with bit lines BL of the memory device. One of the memory cells 10 is located at each intersection of a word line WL and a pair of adjacent bit lines BL and BLB.

[0029] As shown in FIG. 2, the memory cell 10 includes two CMOS inverters 12 and 13 that each have a pull-up PMOS transistor and a pull-down NMOS transistor. The two inverters 12 and 13 are cross-coupled (i.e., for each inverter, the interconnection node for the two transistors of the inverter is connected to the gate terminals of the two transistors of the other inverter). More specifically, the first inverter 12 has a PMOS transistor M3 and an NMOS transistor M1 connected in series, and the second inverter 13 has a PMOS transistor M4 and an NMOS transistor M2 connected in series. The gate terminals of the transistors M1 and M3 of the first inverter 12 are connected to a first node N1, which is the series connection point of the drain terminals of the second inverter 13. Likewise, the gate terminals of the transistors M2 and M4 of the second inverter 13 are connected to a second node N2, which is the series connection point of the drain terminals of the first inverter 12.

[0030] The second node N2 is coupled to a first bit line BL through a first access transistor (e.g., PMOS transistor) M5, the first node N1 is coupled to a second bit line BLB through a second access transistor (e.g., PMOS transistor) M6, and the gate terminals of the access transistors M5 and M6 are connected to the word line WL. The source terminals of the two PMOS transistors M3 and M4 are connected to a voltage supply line Vdd. Further, the source terminals of the two NMOS transistor M1 and M2 of the first and second inverters are connected to a precharge line PL. The precharge line PL is an added line that is common to all the memory cells 10 in a row (i.e., all those connected to one of the word lines). FIG. 3 shows a portion of the structure of a memory matrix that includes such memory cells 10.

[0031] During operation of the SRAM memory device of FIG. 3, when a memory cell is not accessed or read from, the corresponding precharge line PL is coupled to ground. This provides the ground potential to the source terminals of the NMOS transistors M1 and M2 so that both data retention and read operations are performed in the same manner as in a conventional SRAM memory cell. Further, the PMOS access transistors M5 and M6 do not affect the transfer of data from the memory cell to the bit lines. On the other hand, a memory cell is written to using the following process.

[0032] First, a pulse is supplied to the corresponding precharge line PL so as to “reset” the first and second nodes N1 and N2 of each memory cell connected to the precharge line to the high (1) logic level. More specifically, the node N1 or N2 that had been at a low (0) logic level is raised to the voltage level Vdd−Vtn by the corresponding NMOS transistor M1 or M2. Next, the data to be written to the memory cell is placed on the corresponding bit lines BL and BLB as a small voltage difference (i.e., the levels of the bit lines are set according to the memory word to be written). The smaller the voltage difference, the less power is consumed.

[0033] Then, the word line WL is activated so as to supply a negative pulse to the gate terminals of the PMOS access transistors M5 and M6. Because the pulse on the precharge line PL eliminates the current path to ground, the bit lines BL and BLB can easily pass the small voltage difference to the internal nodes N1 and N2. That is, the PMOS access transistors M5 and M6 allow a small voltage difference having a value centered around the supply voltage Vdd to be transferred to the internal nodes. In contrast, conventional memory cells using NMOS access transistors require the voltage difference between the bit lines to be larger than the threshold voltage of an NMOS transistor.

[0034] Next, the pulse to the precharge line PL is ended and the memory cell 10 acts as a latch-type sense amplifier. More specifically, as the potential on the precharge line PL moves back towards ground, the symmetric internal nodes N1 and N2 become unstable and latch in the direction of the small voltage difference that was received from the bit lines. Finally, the word line WL is deactivated to uncouple the internal nodes N1 and N2 of the memory cell from the bit lines and end the write operation.

[0035] FIG. 4 is a timing diagram for such a write operation in the SRAM memory device of FIG. 3. The illustrated write operation relates to the writing of two memory cells of a memory word.

[0036] A memory cell according to a second embodiment of the present invention is shown in FIG. 5. In this embodiment, NMOS transistors are used as the access transistors M5 and M6. As a result, in the second embodiment the bit lines are precharged low between operations, and the internal nodes N1 and N2 are precharged high before a write operation.

[0037] FIG. 6 shows a memory cell according to a third embodiment of the present invention. In the third embodiment, the source terminals of the pull-up transistors M3 and M4 are connected to a precharge line PL that runs parallel to the word line. Additionally, the source terminals of the pull-down transistors M1 and M2 are connected to ground, and NMOS transistors are used as the access transistors M5 and M6. As a result, in the third embodiment the bit lines are precharged low between operations, and the internal nodes N1 and N2 are precharged low before a write operation.

[0038] FIG. 7 shows a memory cell according to a fourth embodiment of the present invention. The fourth embodiment substantially corresponds to the memory cell of the third embodiment (FIG. 6), except that PMOS transistors are used as the access transistors M5 and M6. In particular, the source terminals of the pull-up transistors M3 and M4 are connected to a precharge line PL, a PMOS access transistor M5 is connected between node N2 and one bit line, and another PMOS access transistor M6 is connected between node N1 and the other bit line. The gate terminals of the access transistors M5 and M6 are connected to the word line WL. As a result, in the fourth embodiment the bit lines are precharged high between operations, and the internal nodes N1 and N2 are precharged low before a write operation.

[0039] A dual port memory cell according to a fifth embodiment of the present invention is shown in FIG. 8. The dual port feature provides a memory cell that is connected to a pair of word lines WL_PORT1 and WL_PORT2 and two pairs of bit lines BL_PORT1 and BL_PORT2, and BLB_PORT1, BLB_PORT2. As shown, the first word line WL_PORT1 and the first bit line pair BL_PORT1 and BLB_PORT1 are connected to the memory cell in the same manner as in the first embodiment (FIG. 2). Additionally, a second pair of PMOS access transistors M7 and M8 is provided to selectively couple the cell to the second word line WL_PORT2 and the second bit line pair BL_PORT2 and BLB_PORT2.

[0040] More specifically, the second node N2 is connected to the one of the second bit lines BL_PORT1 through the PMOS access transistor M7, and the first node N1 is connected to the other of the second bit lines BLB_PORT2 through the PMOS access transistor M8. The gate terminals of the added PMOS access transistors M7 and M8 are connected to the second word line WL_PORT2. In this dual port embodiment, simultaneous read and write operations from the two ports PORT1 and PORT2 to the same memory cell are prohibited. In further dual port embodiments, NMOS transistors are used as the access transistors (as in the second embodiment) and/or the source terminals of the pull-up transistors are connected to the precharge line PL (as in the third embodiment).

[0041] As previously explained, the present invention provides a RAM memory cell having cross-coupled inverters with either the pull-down transistors or the pull-up transistors of the inverters being connected to a precharge line. During a write operation, the precharge line is used to “reset” the internal nodes of the memory cell to a predetermined state. This allows data to be stored by creating a small voltage difference between the bit lines. Thus, the memory cell of the present invention allows power consumption to be reduced without introducing significant complexities into the writing process. The present invention is particularly useful for memory devices with very long word lengths.

[0042] While there has been illustrated and described what are presently considered to be the preferred embodiments of the present invention, it will be understood by those skilled in the art that various other modifications may be made, and equivalents may be substituted, without departing from the true scope of the invention. Additionally, many modifications may be made to adapt a particular situation to the teachings of the present invention without departing from the central inventive concept described herein. Furthermore, embodiments of the present invention may not include all of the features described above. Therefore, it is intended that the present invention not be limited to the particular embodiments disclosed, but that the invention include all embodiments falling within the scope of the appended claims.