Semiconductor Memory Device

The present invention generally relates to semiconductor memory devices and, more particularly, to a semiconductor memory device such as a SRAM (static random access memory) which writes and reads information at a high speed.

Recently, various semiconductor memory devices such as the SRAM have been developed as semiconductor memory devices which write and read information at a high speed. The data is held in the SRAM as long as a predetermined power source voltage is maintained; further, no refresh operation is required, as in the case of a DRAM (dynamic random access memory). In the semiconductor memory devices such as the SRAM, there are demands to reduce the power consumption and to increase the operation speed.

As a conventional semiconductor memory device of this type, there is a single port SRAM shown in FIG. 1, for example. In this SRAM, Tn1 and Tn2 represent n-channel MOS transistors, INV1 and INV2 represent inverters, WL represents a word line, and BL and BL represent bit lines.

This SRAM uses the same bit lines BL and BL for writing and reading information. For this reason, the potentials of the bit lines BL and BL are held to the state of the write data after the write operation ends, and it takes almost no time to read out the data immediately after the write operation. However, because the same bit lines BL and BL are used for the write and read operations, there is a need to watch the change in the data (bit error) held by the memory cell. This bit error is a phenomenon in which the data held in the memory cell is inverted due to the load of the bit lines BL and BL when the word line WL is switched at the time of the data read out.

That is, in the SRAM of the type shown in FIG. 1, the word line WL and the bit lines BL and BL are respectively made up of several tens to several hundreds of aluminum lines, and the capacitance of the bit lines BL and BL is considerably large. For this reason, when the word line WL is switched, the potential difference of the bit lines BL and BL has a greater effect on the potentials of the n-channel MOS transistors Tn1 and Tn2 than the inverters INV1 and INV2 which invert the potentials of the n-channel MOS transistors Tn1 and Tn2, and there are cases where the above bit error occurs.

Conventionally, measures were taken against the bit error such as (1) reducing the potential difference of the bit lines BL and BL by pulling up the lower voltage level of the voltage levels applied to the bit lines BL and BL to approximately 2 V, for example, and (2) short-circuiting the bit lines BL and BL before switching the word line WL so as to eliminate the potential difference of the bit lines BL and BL.

However, according to the conventional method (1), the lower voltage level of the bit lines BL and BL is pulled up to approximately 2 V. As a result, a voltage is applied to the bit lines even when the voltage level corresponds to the low level, presenting problem of the power consumption increasing.

On the other hand, according to the conventional method (2), the bit lines BL and BL are short-circuited before switching the word line WL. For this reason, a signal for short-circuiting the bit lines BL and BL must be generated internally. Hence, even if an attempt is made to perform a read operation after the write operation ends, for example, a time margin is required to generate this signal, and there is a problem in that the operation speed deteriorates.

As a measure against the the above two problems, a SRAM shown in FIG. 2 has been proposed. In this SRAM, Tn3 and Tn4 represent n-channel MOS transistors, INV3 and INV4 represent inverters, WBL and WBL represent bit lines exclusively for writing, and RBL and RBL represent bit lines exclusively for reading.

In this SRAM, dividing the bit lines BL and BL into the bit lines WBL and WBL exclusively for reading and the bit lines RBL and RBL exclusively for reading eliminates the time margin required by the above conventional method (2) in order to internally generate the signal for short-circuiting the bit lines BL and BL. However, according to this structure in which the bit lines BL and BL are divided into the bit lines WBL and WBL exclusively for reading and the bit lines RBL and RBL exclusively for reading, a problem occurs if the capacitance of the bit lines BL and BL is large, that is, in the case of a large scale SRAM. In other words, when a write enable signal is changed from a low level to a high level and the read operation is carried out immediately after the write operation, the potentials of the n-channel MOS transistors Tn3 and Tn4 are inverted by the inverters INV3 and INV4 after the bit line potential on the write side is inverted when switching the word line WL, and it takes a considerably long time for the bit line potential on the read side to be inverted. As a result, when the data is read out immediately after the write operation, there are cases where the data before the write operation is carried out is read out, that is, the previous data before the potentials of the n-channel MOS transistors Tn3 and Tn4 are inverted is read out.

In FIG. 3, (a) shows an address signal, (b) shows the write enable signal, (c) shows the bit line potential on the write side, (d) shows the bit line potential on the read side, and (e) shows the output data of the memory.

Accordingly, a predetermined time interval must be provided before the read operation which is carried out immediately after the write operation, and there was a problem in that the access time becomes extremely long.

Accordingly, it is a general object of the present invention to provide a novel and useful semiconductor memory device in which the problems described above are eliminated.

Another and more specific object of the present invention is to provide a semiconductor memory device comprising a memory part, two bit lines exclusively for writing which are coupled to the memory part and used when writing information into the memory part, two bit lines exclusively for reading which are coupled to the memory part and used for reading information held in the memory part, and a short-circuiting circuit coupled to the bit lines exclusively for reading, which short-circuits the two bit lines exclusively for reading and sets them to the same potential in a write mode in which information is written into the memory part via the bit lines exclusively for writing. According to the present invention, the read out from the same address immediately after the write operation can also be carried out at a high speed, and a high-speed memory access is constantly possible while suppressing the power consumption.

Other objects and further features of the present invention will be apparent from the following detailed description when read in conjunction with the accompanying drawings.

FIG. 1 is a circuit diagram showing an essential part of an example of a conventional semiconductor memory device;

FIG. 2 is a circuit diagram showing an essential part of another example of the conventional semiconductor memory device;

FIG. 3 is a timing chart for explaining the operation of the semiconductor memory device shown in FIG. 2;

FIG. 4 is a circuit diagram showing an essential part of an embodiment of a semiconductor memory device according to the present invention;

FIG. 5 is a block diagram showing the entire construction of the embodiment of the present invention;

FIG. 6 is a timing chart for explaining a read operation of the embodiment;

FIG. 7 is a timing chart for explaining a write operation of the embodiment;

FIG. 8 is a timing chart for explaining a normal read operation of the embodiment;

FIG. 9 is a timing chart for explaining a read operation immediately after a write operation in the conventional example shown in FIG. 2; and

FIG. 10 is a timing chart for explaining a read operation immediately after a write operation in the embodiment.

FIG. 4 shows an embodiment of a semiconductor memory device according to the present invention.

A SRAM shown in FIG. 4 is provided with a flip-flop FF having respective input and output ends of two inverters INV1 and INV2 respectively cross connected, two switching transistors Tn1 and Tn2 respectively connected to two output ends of the flip-flop FF and two write bit lines WBL and WBL exclusively for writing and having gates connected to a write word line WWL exclusively for writing, two inverters INV3 and INV4 for inverting outputs from the two output ends of the flip-flop FF, and two switching transistors Tn3 and Tn4 respectively connected to the inverters INV3 and INV4 and two read bit lines RBL and RBL exclusively for reading and having gates connected to a read word line RWL exclusively for reading. When writing data to the flip-flop FF from the write bit lines WBL and WBL, the two read bit lines RBL and RBL are short-circuited so as to set the read bit lines RBL and RBL to the same potential.

In FIG. 4, 1 represents a sense amplifier circuit, 2 represents a write amplifier circuit, Tr1 and Tr2 respectively represent switching transistors for reading, Tw1 and Tw2 respectively represent switching transistors for writing, and Ts represents an n-channel MOS transistor forming a short circuit 50. In addition, a precharge pulse which will be described later is applied to transistors Tpc1 through Tpc3.

In this embodiment, the two bit lines exclusively for reading are short-circuited when writing data to the flip-flop from the bit lines exclusively for writing, and the two bit lines exclusively for reading are set to the same potential.

In other words, even when carrying out the read operation immediately after the write operation, there is no need to invert the potentials of the bit lines exclusively for reading from the power source potential and the ground potential, and it is possible to quickly read the data by cancelling the short-circuiting of the bit lines exclusively for reading after the write operation ends.

FIG. 5 is a block diagram showing the entire construction of this embodiment. In FIG. 5, the SRAM generally comprises a sense amplifier 1, a write amplifier 2, a column selector 3, a memory cell array 4, an address buffer 5, a column decoder 6, an address transition detection (hereinafter referred to as ATD) circuit 7, a row decoder 8, an enable buffer 9 and a precharger 10. RE denotes a read enable signal and WE denotes a write enable signal.

The sense amplifier 1 detects and amplifies a signal from the memory cell at the time of the read out, and the write amplifier 2 supplies a write signal to the memory cell. The column selector 3 determines a signal path of the write and read signals, and is made up of the switching transistors Tr1, Tr2, Tw1 and Tw2 shown in FIG. 4. The memory cell array 4 has the memory cells for storing information arranged in a matrix form. The address buffer 5 latches address information for carrying out the write and read operations, and outputs this information to the column decoder 6, the ATD circuit 7 and the row decoder 8 together with an enable signal from the enable buffer 9.

The ATD circuit 7 detects an address transition and generates a predetermined pulse. The generated pulse is output to the memory cell array 4 via the precharger 10 as a precharge pulse.

A diagram showing the particular circuit construction is similar to the circuit diagram shown in FIG. 4. Next, a description will be given of the read and write operations by referring to FIGS. 6 and 7.

In FIGS. 6 and 7, t.sub.RC denotes a read cycle time, t.sub.AA denotes an address access time, t.sub.OH denotes a data hold time corresponding to the address transition, t.sub.WC denotes a write cycle time, t.sub.AW denotes an address settle time corresponding to the write enable signal WE, t.sub.AS denotes an address setup time t.sub.WP denotes a write pulse width t.sub.WR denotes a write recovery time, t.sub.DW denotes a data setup time, t.sub.DH denotes a data hold time, and VALID denotes a time period in which the data is valid.

First, in the case of the read operation, the data is output a predetermined time (address access time t.sub.AA) after the address undergoes a transition, as shown in FIG. 6, and the level of the write enable signal WE becomes high.

In FIG. 6, (a) shows the address signal and (b) shows the output data of the memory.

Next, in the case of the write operation, when the level of the write enable signal WE is made low after the address settles, as shown in FIG. 7, the n-channel MOS transistor Ts, which forms the short-circuiting circuit 50 turns OFF and the state becomes the write state. The write pulse width t.sub.WP indicates a minimum value, and the write operation is not guaranteed if the width of the low-level write enable signal WE is less than or equal to the write pulse width t.sub.WP. In addition, the set up time t.sub.DW of the input data and the hold time t.sub.DH are determined with respect to the rising edge of the write enable signal WE.

In FIG. 7, (a) shows the address signal, (b) shows the write enable signal, and (c) shows the output data of the memory.

That is, if the read operation is carried out immediately after the write operation, the reading of the data is quickly carried out after the write operation ends by cancelling the short-circuiting of the bit lines exclusively for the reading.

Accordingly, the read out from the same address after the write operation ends is carried out at a high speed even with the low power consumption, and the characteristic of the RAM is improved.

In this embodiment, when carrying out the read operation immediately after the write operation, there is no need to invert the respective potentials of the bit lines used exclusively for reading from the power source potential and the ground potential. Thus, it is possible to quickly read out the data after the write operation ends by cancelling the short-circuiting of the bit lines used exclusively for reading, and a high-speed processing is possible while nevertheless suppressing power consumption.

Next, a description will be given of the effects of this embodiment in comparison with the conventional example shown in FIG. 2.

First, the normal read operation will be described in conjunction with FIG. 8. If the level of the precharge pulse becomes high as shown in FIG. 8(b) after the address signal undergoes a transition as shown in FIG. 8(a), the read bit lines RBL and RBL in FIG. 4 are short-circuited because the transistor T.sub.TPC2 is ON. Hence, the potentials of the read bit lines RBL and RBL become the same as shown in FIG. 8(c), and the output data can be obtained as shown in FIG. 8(d) by amplifying the potential difference of the read bit lines RBL and RBL thereafter by the sense amplifier circuit 1. Accordingly, the read out of the data is carried out at a high speed. This normal read operation is the same as that of the conventional example shown in FIG. 2.

FIG. 9 is a timing chart for explaining the case where the read operation is carried out immediately after the write operation in the conventional example shown in FIG. 2. In FIG. 9, (a) shows the address signal, (b) shows the precharge pulse, (c) shows the write enable signal, (d) shows the data of the memory cell, (e) shows the potentials of the read bit lines RBL and RBL, and (f) shows the output data of the memory. In addition, a1, a1', b1 and b1' respectively correspond to the potentials at the nodes shown in FIG. 2.

In the case of FIG. 9, after the level of the write enable signal shown in (c) becomes low and the data of the memory cell inverts as shown in (d), a time ta required to obtain the output data of (f) measured from the time when the data of the memory cell is read out to the read bit lines RBL and RBL as shown in (e) to the time when the potential difference is amplified in the sense amplifier circuit 1, is long. For this reason, according to the conventional example shown in FIG. 2, the read operation immediately after the write operation cannot be carried out at a high speed although the normal read operation can be carried out at the high speed. As a result, the access time is long when the read operation is carried out with respect to the same address immediately after the write operation.

On the other hand, FIG. 10 is a timing chart for explaining the read operation immediately after the write operation in the embodiment shown in FIG. 4. In FIG. 10, (a) shows the address signal, (b) shows the precharge pulse, (c) shows the write enable signal, (d) shows the data of the memory cell, (e) shows the potentials of the read bit lines RBL and RBL, and (f) shows the output data of the memory. In addition, a1, a1', b1 and b1' respectively correspond to the potentials at the nodes shown in FIG. 4.

In the case of FIG. 10, after the level of the write enable signal shown in (c) becomes low and the data of the memory cell inverts as shown in (d), a time tb required to obtain the output data of (f), measured from the time when the data of the memory cell is read out to the read bit lines RBL and RBL shown as in (e) to the time when the potential difference is amplified in the sense amplifier circuit 1, is extremely short. This is because the transistor Ts of the short-circuiting circuit 50 short-circuits the read bit lines RBL and RBL and sets them to the same potential in response to the low-level write enable signal. As a result, the data of the memory cell is read out to the read bit lines RBL and RBL at a higher speed in FIG. 10, when compared to the case shown in FIG. 9, and it is possible to obtain the output data extremely quickly. Therefore, it may be seen that even the read operation immediately after the write operation can also be carried out at the high speed according to this embodiment.

Further, the present invention is not limited to these embodiments, but various variations and modifications may be made without departing from the scope of the present invention.

As described above, according to the semiconductor memory device of the present invention, a short-circuiting circuit is provided to short-circuit the bit lines which are used exclusively for reading, when writing data so as to set these bit lines to the same potential, and thus, it is possible to carry out a high speed read operation even with respect to the same address, immediately after the write operation and further, and it is possible to constantly realize a high-speed memory access while suppressing the power consumption. For these reasons, the present invention is extremely useful from the practical point of view.