This application is based upon and claims the benefit of priority from prior Japanese Patent Application No. 2010-267829, filed Nov. 30, 2010, the entire contents of which are incorporated herein by reference.
Embodiments described herein relate generally to memory systems, and is applied to, for example, a semiconductor memory system including multiple types of memories integrated into a single chip.
An example of a semiconductor memory system including multiple types of memories integrated into a single chip is a semiconductor memory system including a NAND flash memory (memory unit) and a SRAM (Static Random Access Memory) integrated in a single chip (see Japanese Patent Application Publication No. 2006-73141).
In general, according to one embodiment, a memory system includes (1) a memory cell array including a plurality of memory cells electrically connected to pairs of bit lines when a word line is activated; latch portions connected to respective pairs of bit lines; a sense amplifier connected to the latch portions; and a control circuit configured to control the latch portions for a reading operation so that data in all memory cells connected to the word line, once selected, come to be held in the corresponding latch portions as a group.
Referring to the drawings, description will be provided for a first embodiment. For the convenience of explanation, the same portions will be denoted by the same reference signs throughout all the drawings. In addition, dimensional ratios among portions are not limited to those indicated in the drawings.
A memory system of the first embodiment will be described with reference to
As shown in
First of all, the NAND flash memory 2 will be described by use of
The NAND flash memory 2 functions as a main memory unit of the memory system 1. As shown in
As shown in
Multiple bit lines BL0 to BLm (“m” is natural number) are arranged, extending in a direction in which the NAND strings NS extend (i.e., in a first direction), over the NAND strings above a semiconductor substrate. The multiple bit lines BL0 to BLm are electrically connected to end portions of the NAND strings NS, respectively.
On the other hand, multiple word lines WL0 to WL31 are arranged, extending in a direction (second direction) orthogonal to the first direction, side-by-side at predetermined intervals in the first direction. In this respect, the first direction is concurrently a direction in which active regions extend.
Multiple selection gate lines SGS and SGD are arranged in parallel outside the respective word lines WL0 and WL31 with the multiple word lines WL0 to WL31 interposed in between.
Each NAND string NS includes multiple memory cells MT0 to MT31, and first and second selection gate transistors ST1 and ST2. Each memory cell MT has a stacked gate structure which includes: a charge storage layer formed above the semiconductor substrate with a gate insulating film interposed in between; and a control gate formed above the charge storage layer with an inter-gate insulating film interposed in between. Incidentally, the number of memory cells MT is not limited to 32, and may be any one of 8, 16, 34, 128, 256, etc. No specific restriction is imposed on the number of memory cells MT. In addition, each memory cell transistor MT may have a MONOS (Metal Oxide Nitride Oxide Silicon) structure that is obtained by a method of trapping electrons in a nitride film, instead of the stacked gate structure.
Each multiple memory cell MT0 to MT31 is formed in a portion corresponding to an intersection between the word lines WL and the corresponding bit line BL, and the memory cells are connected together in series in the direction in which the active regions extend (i.e., in the first direction).
In addition, as shown in
As shown in
The multiple NAND strings NS are formed in a matrix in the memory cell array 10. Each set of memory cells MT sharing the same word line WL throughout all the NAND strings NS constitutes a page, which is a data reading/writing unit. Furthermore, each set of multiple NAND strings NS sharing the same word line WL constitutes a block, which is a data erasing unit.
The page buffer 12 is capable of holding one page of data. During a data read operation, the page buffer 12 temporarily holds data that is read from the memory cell array 10, and transfers the data to the RAM unit 3. In addition, during a data write operation, the page buffer 12 temporarily holds data to be written from the RAM unit 3, and transfers the data to the memory cell array 10.
A region in the page buffer 12 is used to hold main data, and the remaining region in the page buffer 12 is used to hold the parity, etc.
The row decoder 11 selects a desired word line (s) WL in the memory cell array 10. In addition, the column decoder selects a desired column (s), namely, a desired bit line (s) BL in the memory cell array 10.
The voltage generating circuit 13 generates a voltage needed to program, read, or erase data by raising or lowering a voltage given from the outside. Thus, the voltage generating circuit 13 supplies the generated voltage to the row decoder 11, for example. Hence, the voltage generated by the voltage generating circuit 13 is applied to a word line(s) WL.
<<Sequencer>>
The sequencer 14 controls the operation of the NAND flash memory 2 as a whole. Once receiving a NAND interface command (“NAND I/F command”) from the control unit 4, the sequencer 14 executes a sequence corresponding to the NAND interface command (for example, a sequence for programming data). In accordance with the sequence, the sequencer 14 controls the operation of the page buffer 12, the operation of the voltage generating circuit 13, etc. The sequencer 14 operates in synchronism with an internal clock ICLK transferred to the sequencer 14 from the oscillator 15, which will be described below.
<<Oscillators>>
The oscillator 15 (clock generator) generates the internal clock ICLK. The oscillator 15 transfers the generated internal clock ICLK to the sequencer 14.
The oscillator 16 (clock generator) generates the other internal clock ACLK. The oscillator 16 transfers the generated internal clock ACLK to the control unit 4, etc. The internal clock ACLK is a clock serving as a reference with which the controller 4 and the like operate in synchronism.
As shown in
During the data read operation, the ECC unit 20 detects and corrects an errors included in data that is read from the NAND flash memory 10. On the other hand, during the data write operation, the ECC unit 20 generates a parity for data that needs to be programmed.
The ECC unit 20 includes an ECC buffer 21 and an ECC engine 22. The ECC buffer 21 is connected to the page buffer 12 of the NAND flash memory 10 via the NAND data bus. Further, the ECC buffer 21 is connected to the SRAM unit 30 via the ECC data bus.
During the data read operation, the ECC buffer 21 holds data that is transferred from the page buffer 12, and transfers data, which finishes an ECC process (which finishes error correction during the data load operation), to the SRAM unit 30. On the other hand, during the data write operation, the ECC buffer 21 holds data that is transferred from the SRAM unit 30, and transfers data and the corresponding parity, which are transferred from the SRAM unit 30.
The ECC engine 22 performs an ECC process by use of data held in the ECC buffer 21. The ECC engine 22 employs, for example, a one-bit correction method using a Hamming code. In addition, the ECC engine 22 uses the minimum parity data needed for the correction process.
<<SRAM Unit>>
As shown in
The multiple data RAM has multiple banks. Each bank has multiple SRAM memory cells in it. Word lines (for example, 32 word lines) connected to the SRAM memory cells are connected to the row decoder 34. Furthermore, pairs of bit lines (for example, 256 pairs of bit lines) connected to the SRAM memory cells are connected to the sense amplifier unit 33.
Each sense amplifier unit 33 includes multiple sense amplifiers. In a case where, as illustrated in
As in the memory cell array 10, the memory cell array 32 of each data RAM includes a region in which to hold main data and the other region in which to hold the parity, etc.
The sense amplifier unit 33 of each data RAM senses and amplifies data that is read from the SRAM cells to the pairs of bit lines BL, /BL. The row decoder 34 selects some out of the word lines WL of the memory cell array 32 in each data RAM.
<<Configuration between SRAM 30 and Interface Unit 40>>
Descriptions will be provided for a configuration between the SRAM unit 30 and the interface unit 40 by use of the example shown in
In a case where, as shown in
In each bank, the SRAM memory cells are connected to the sense amplifiers S/A1 to S/A16. An address is set for each bank. In the case shown in
As shown in
The data latch selector and the burst selector are controlled by receiving a selector address signal for determining which address should be selected, and a clock, from a burst read control circuit.
Each master latch and a slave latch are capable of holding a page of data. When a clock is inputted into the slave latch from the bust read control circuit, data is outputted from the slave latch to the interface 43.
Next, using
Each SRAM memory cell has a configuration as shown in
In addition, an equalizer line /EQL is commonly connected to each pair of bit lines BL, /BL in each bank in the SRAM memory cell array. Bit line pre-charging transistors (P-channel MOS transistors) P3, P4 for pre-charging the potentials of the pairs of bit lines BL, /BL by use of the power supply VDD and an equalizer-dedicated transistor (P-channel MOS transistor) P5 are provided between intersections between the equalizer line /EQL and each pair of bit lines BL, /BL.
Moreover, a latch circuit (a latch portion) is connected to each pair of bit lines BL, /BL in the SRAM memory cell array. The latch circuit has a configuration as follows.
The latch circuit has the configuration which is shown in
Moreover, a pair of transfer gates are formed in each pair of bit lines BL, /BL. A necessary pair of bit lines BL, /BL are selected from the 16 pairs of bit lines BL, /BL connected to each sense amplifier by use of its corresponding pair of transfer gates. To put it specifically, an access controller 50 inputs an internal control signal CSL into the gate of the PMOS transistor in the transfer gate connected to the bit line BL, and inputs an internal control signal /CSL into the gate of the other PMOS transistor in the transfer gate connected to the bit line /BL. The access controller 50 inputs “H” as the internal control signal CSL and “L” as the internal control signal /CSL into a selected pair of bit lines BL, /BL. On the other hand, the access controller 50 inputs “L” as the internal control signal CSL and “H” as the internal control signal /CSL into the other unselected pairs of bit lines BL, /BL.
The boot RAM temporarily holds a boot code for activating the memory system 1, for example. The DQ buffer 31 temporarily holds data when the data is written into the data RAMs, or when the data is read from the data RAMs.
As shown in
In addition, use of the RAM/register bus enables data to be transmitted between the DQ buffer 31 and the burst buffers (the burst buffers shown in
The interface unit 40 includes the burst buffers 41, 42, and an interface (an I/F shown in
The burst buffers 41, 42 are electrically connected to the DQ buffer 31 and the controller unit 4 via the RAM/Register bus. As a result, data can be transferred among the DQ buffer 31, the controller unit 4, and each of the burst buffers 41, 42.
The burst buffers 41, 42 are electrically connected to the interface 43 via a DIN/OUT bus. As a result, data can be transferred between the interface 43 and each of the burst buffers 41, 42. The burst buffers 41, 42 temporarily hold data that is given to the burst buffers 41, 42 from a host apparatus via the interface 43, or data that is given to the burst buffers 41, 42 from the DQ buffer 31.
The interface 43 can be connected to the host apparatus outside the memory system 1. The interface 43 controls the input and output of various signals, such as data, control signals, and addresses, to and from the host apparatus.
Examples of the signals include: a chip enable signal /CE for enabling the entire memory system 1, an address valid signal /AVD for latching an address, a clock CLK for a burst read, a write enable signal /WE for enabling a write operation, and an output enable signal /OE for enabling the output of data to the outside.
The interface 43 is electrically connected to the burst buffer 41, 42 via the DIN/OUT bus. The interface 43 transfers control signals from the host apparatus concerning a data read request, a load request, a write request, etc. to an access controller 50. For a data read operation, the interface 43 outputs data in the burst buffers 41, 42 to the host apparatus. For a data write operation, the interface 43 transfers data, which is given to the interface 43 from the host apparatus, to the burst buffers 41, 42.
The access controller 50 receives a control signal and an address from the interface 43. In response, the access controller 50 controls the SRAM 30 and the control unit 4 in order for an operation, which satisfies a request of the host apparatus, to be executed.
To put it specifically, in response to the request from the host apparatus, the access controller 50 puts either the SRAM 30 or a register 60 inside the controller unit 4 in an active state. Subsequently, the access controller 50 issues a write command or a read command of data (denoted by reference sign Write/Read in
As shown in
The register 60 sets up an operational status of a function. The register 60 allocates part of an external address space to this end. Thereby, the external host apparatus reads or writes either an address signal or a control signal, such as a command, from and to the allocated part of the external address space of the register 60 via the interface 43.
Once the address signal or the control signal, such as a command, is written into the predetermined part of the external address space of the register 60, the CUI 61 recognizes that a function execution command is given to the CUI 61, and issues an internal command signal.
Upon reception of a command issued from the address/command generator circuit 63, which will be described later, or the internal command signal from the CUI 61, the state machine 62 controls an internal sequence operation, depending on the type of command.
The address/command generator circuit 63 generates an address signal and a control signal, such as a command, to the NAND flash memory 2 as necessary during the internal sequence operation.
The address/timing generator circuit 64 generates an address and a control signal, such as a signal representing timing, for controlling the SRAM 30, as necessary during the internal sequence operation.
Next, as part of a method of operating the memory system of the first embodiment, the operation of the memory system until data in the SRAM memory cells connected to one word line WL inside the bank 0 shown in
First, in step S1, upon reception of a command from the interface 43, the access controller 50 charges all the pairs of bit lines BL, /BL inside the bank 0. For example, the access controller 50 charges the pairs of bit lines BL, /BL by making control in order that the transistors connected to MDQ and /MDQ, as shown in
Subsequently, in step S2, the access controller 50 controls the SRAM 30, and puts the equalizer line /EQL0 of the bank 0 into “H” once a clock CLK is inputted into the bank 0 via the row decoder 34. In response, the P-channel transistors connected to the equalizer line /EQL turn into an “OFF” state, and the pairs of bit lines BL, /BL enter a floating state.
In step S3, the access controller 50 controls the SRAM 30, and puts a word line WL, which is selected in the bank 0 via the row decoder 34, into “H.” Thereby, data of all the memory cells connected to the selected word line WL (for example, 256-bit data) is transferred to the pairs of bit lines BL, /BL in the floating state. Thereafter, as shown in
The internal control signal FS is a signal inputted to confirm data. The internal control signal FS is designed to be inputted when the electric potential difference between paired bit lines BL, /BL exceeds a predetermined electric potential difference.
Let us assume that, for example, “1” is held in one SRAM memory cell. In other words, let us assume that the node K1 of one SRAM memory cell is put in “L” while the node K2 of the same SRAM memory cell is put in “H.” In this case, the MOS transistor P1 is off, the MOS transistor P2 is on, the MOS transistor N1 is on, and the MOS transistor N2 is off. Once a word line WL is connected to this SRAM memory cell, “H” is inputted into the gates of the respective N-channel MOS transistors N3, N4. Thereby, the N-channel MOS transistors N3, N4 are put into an “ON” state. When the word line WL is selected, the SRAM memory cell is electrically connected to the corresponding latch circuit. Thus, the bit line BL is kept charged, while the bit line /BL is discharged. Once the electric potential difference between the pair of bit lines BL, /BL exceeds the predetermined electric potential difference, data is confirmed by an internal control signal FS. Thereafter, the access controller 50 controls the SRAM 30, and inputs an internal control signal SEN, which is put in an “H” state, into the latch circuit. Thus, the access controller 50 causes the data to be held in the latch circuit (step S5). The data held in the pair of bit lines BL, /BL come to be held in the latch circuit. The data can be held by keeping the sources of the respective N-channel MOS transistors N5, N6 at the ground potential with the internal control signal SEN put in the “H” state. In other words, the data held in the nodes K1, K2 are transferred to the nodes K3, K4 in the latch circuit. Thereby, the data held in the SRAM memory cell comes to be held in the latch circuit. After data of all the memory cells connected to the word line WL are held in the corresponding latch circuits, the internal control signal FS is put into an “L” state, and the word line WL is put into “L.” (see
In step S6, the access controller 50 controls the SRAM 30, and selects a pair of bit lines BL, /BL by inputting a desired internal control signal into a pair of transfer gates CSL, /CSL connected to the pair of bit lines BL, /BL inside the bank 0, while keeping the internal control signal SEN in the “H” state. In other words, in the case shown in
In this regard, the pairs of bit lines BL, /BL are sequentially selected with the internal control signal SEN kept in “H,” until all the data held in the SRAM memory cells connected to the selected word line WL is transferred to the burst buffers 41, 42. For example, in order to hold 256-bit data in the latch circuits in the bank 0, the selection and non-selection of the pairs of bit lines BL, /BL are carried out 16 times.
Thereby, the data held in the SRAM memory cells connected to one word line WL in the bank 0 can be read to the corresponding sense amplifier S/A.
In the embodiment shown in
By employing the foregoing configuration and operations, the first embodiment can provide a memory system configured to reduce the power consumption while reading data from the SRAM memory unit. Detailed descriptions will be hereinbelow provided.
The memory system of the first embodiment sequentially selects necessary pairs of bit lines BL, /BL while keeping the internal control signal SEN of the corresponding latch circuits in “H.” Thereby, the memory system is configured to read data held in the latch circuits corresponding to all the memory cells connected to a selected word line WL. For example, in order to hold 256-bit data in latch circuits in the bank 0, the memory system carries out 16 times of selection and non-selection of the pairs of bit lines BL, /BL. In this process, the memory system is configured to read the data from all the memory cells connected to one selected word line WL with the word line electrically charged only once, instead of repeatedly charging and discharging the word line WL each time the selection and non-selection are switched among the pairs of bit lines BL, /BL. As a result, the memory system no longer needs to recharge the word line WL repeatedly while reading the data. Accordingly, the memory system can reduce the power consumption. In addition, because the memory system of the first embodiment sequentially selects necessary pairs of bit lines BL, /BL while keeping the internal control signal SEN of the corresponding latch circuits in “H,” the memory system is configured to read the data without carrying out equalization each time the selection and non-selection of the pairs of bit lines BL, /BL are changed. As a result, the memory system no longer needs to recharge the word line WL repeatedly while reading the data. Accordingly, the memory system can reduce the power consumption.
Thus, the memory system of the first embodiment can reduce the power consumption while reading the data from the SRAM memory unit.
The memory system of the first embodiment puts the selected word line WL into “L,” once the memory system causes the data held in all the memory cells connected to a selected word line WL to be held in the corresponding latch circuits. Furthermore, the memory system puts the internal control signal SEN into “H” when reading the data from the latch circuits. As opposed to the memory system of the first embodiment, a memory system of Modification 1 is designed to keep the word line WL in “H” and the internal control signal SEN in “L” until the memory system finishes reading the data from all the memory cells connected to the selected word line WL to the burst buffers 41, 42.
The memory system of the first embodiment reads data, which is held in the banks 0 to 3, to the external host apparatus in the sequence from the bank 0, the bank 1, the bank 2, and the bank 3. As opposed to the memory system of the first embodiment, a memory system of Modification 2 transfers data from a certain bank 0, and thereafter transfers data from another bank, the output terminal of whose sense amplifier unit is not commonly connected to the output terminal of the sense amplifier unit 33 of the certain bank 0, when the memory system has multiple sets each consisting of two banks, the output terminals of whose sense amplifier units 33 are commonly connected together. For example, the memory system has two sets of banks, as shown in
The sequential reading of data from the bank 0 and the bank 2, the output terminals of whose sense amplifiers units 33 are not commonly connected together, enables the memory system to charge the pairs of bit lines BL, /BL selected for reading data from the bank 2, while reading data from the bank 0. This makes it possible for the memory system of Modification 2 to read data faster than the memory system of the first example.
It should be noted that the first embodiment allows the word lines WL to be discharged at any time before the equalizer line /EQL is discharged and allows the timing of charging the word lines to be changed as necessary.
While certain embodiments have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel embodiments described herein may be embodied in a variety of other forms. Furthermore, various omissions, substitutions and changes in the form of the embodiments described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.
Number | Date | Country | Kind |
---|---|---|---|
P2010-267829 | Nov 2010 | JP | national |