Patent Grant
7796429
This application is based upon and claims the benefit of priority from prior Japanese Patent Application No. 2007-050390, filed Feb. 28, 2007, the entire contents of which are incorporated herein by reference.
1. Field of the Invention
The present invention relates to a memory controller which controls a semiconductor storage device and a semiconductor device provided with the memory controller.
2. Description of the Related Art
Recently, due to the rapid spread of digital cameras and portable audio players, demand for large-capacity nonvolatile semiconductor memories has increased. A NAND type flash memory (hereinafter sometimes simply referred to as flash memory) is widely used as the nonvolatile semiconductor memory.
In the NAND type flash memory, when the number of memory cells into which electrons are injected is increased, a so-called program disturb is generated. Program disturb is a phenomenon in which data retained by a non-selected memory cell is destroyed. A program disturb is also generated if a series of identical pieces of data exist in a cell array, due to GIDL (Gate Induced Drain Leakage).
In order to prevent the program disturb from occurring, for example, JP-A H7-334991 (KOKAI) proposes a technique of averaging a charge amount in the memory cell array as a whole. However, in the technique, the charge amount is insufficiently averaged, and the generation of a series of identical pieces of data cannot be prevented. Accordingly, unfortunately, the operational reliability of the NAND type flash memory is lowered with the increased capacity of the NAND type flash memory.
A memory controller which controls a semiconductor storage device including nonvolatile memory cells according to an aspect of the present invention includes:
a generating circuit which generates a plurality of first data based on a second data to be collectively written in the memory cells, the second data having multi-bit; and
a selection circuit which retains a cumulative value whose each digit is a cumulative result in each bit of data which is already written in the memory cells, the selection circuit selecting one of the plurality of first data, a selected first data having a better average of digits in a sum of each bit of the selected first data and each digit of the cumulative value than an unselected first data, the selection circuit retaining the sum concerning the selected first data as the new cumulative value.
A semiconductor device according to an aspect of the present invention includes:
a memory controller described above; and
the semiconductor storage device which includes a first memory region and a second memory region, the memory cells being arranged in each of the first memory region and the second memory region, data corresponding to the first data selected by the selection circuit of the memory controller being written in the memory cells of the first memory region, the information indicating the selected first data being written in the memory cells of the second memory region.
A semiconductor device according to a first embodiment of the invention will be described with reference to
A configuration of the NAND type flash memory 10 will be described below. Hereinafter, sometimes the NAND type flash memory 10 is simply referred to as flash memory 10. The flash memory 10 includes a memory cell array 11, a page buffer 12, and a sense amplifier 13. The memory cell array 11 includes a plurality of memory blocks 14, and a plurality of nonvolatile memory cells MC are arranged in each of the memory blocks 14. Each of the memory cells MC is capable of retaining one bit, i.e., binary data. All the memory cells MC included in the same memory block 14 are collectively erased, and pieces of data are collectively written in the plurality of memory cells MC. Hereinafter, the unit of a memory cell MC in which the pieces of data are collectively written is referred to as a page. The page buffer 12 retains pieces of write data to be written in the memory cell MC in page, and the page buffer 12 writes the retained write data in the memory cell MC. The sense amplifier 13 senses and amplifies data read from the memory cell MC.
A configuration of the memory block 14 will be described below with reference to
The regular cell array 15 and the ECC cell array 16 include a plurality of NAND strings. Each of the NAND strings includes selection transistors ST1 and ST2 and 32 memory cell transistors MT. The 32 memory cell transistors MT included in the NAND string are illustrated by way of example only. For example, 8, 32, 64, or 128 memory cell transistors MT may be included in the NAND string. In the selection transistor ST1 included in each NAND string, a drain is connected to one of bit lines BL0 to BL4313 and a gate is commonly connected to a select gate line SGD. In the selection transistor ST2, a source is commonly connected to a source line SL and a gate is commonly connected to a select gate line SGS.
Each of the memory cell transistors MT is a MOS transistor including a stacked gate which is formed on a semiconductor substrate with a gate insulating film interposed therebetween. The stacked gate includes a charge accumulation layer (for example, floating gate) formed on the gate insulating film and a control gate which is formed on the charge accumulation layer with an inter-gate insulating film interposed therebetween. In each NAND string, current paths of the 32 memory cell transistors MT are connected in series between the source of the selection transistor ST1 and the drain of the selection transistor ST2. Control gates of the memory cell transistors MT are sequentially connected to word lines WL0 to WL31 from the memory cell transistor MT located closest to the source side. Accordingly, the drain of the memory cell transistor MT connected to the word line WL31 is connected to the source of the selection transistor ST1, and the source of the memory cell transistor MT connected to the word line WL0 is connected to the drain of the selection transistor ST2.
The 4096 NAND strings having the above-described configuration are arranged in the regular cell array 15 and the 218 NAND strings are arranged in the ECC cell array 16. The 4096 bit lines BL0 to BL4095 are connected to the NAND strings in the regular cell array 15 respectively, and the 218 bit lines BL4096 to BL4313 are connected to the NAND strings in the ECC cell array 16 respectively. Accordingly, the one regular cell array 15 has a 4 k-bit memory capacity per one word line, and the one ECC cell array 16 has a 218-bit memory capacity per one word line.
Eight combinations of the regular cell array 15 and ECC cell array 16 having the above-described configurations are arranged in the one memory block 14. Accordingly, in the memory block 14, the regular cell array 15 has the memory capacity of (4 k bits×8)=4 k bytes per one word line, and the ECC cell array 16 has the memory capacity of (218 bits×8)=218 bytes. In each memory block 14, pieces of data are collectively written in the memory cell transistors MT connected to the same word line. That is, a one-page capacity becomes (4 k bytes+218 bytes). The pieces of data are collectively erased in the memory cell transistors MT in each memory block 14. That is, an erase size becomes ((4 k bytes+218 bytes)×32)=(128 k bytes+6976 bytes) because the number of word lines is 32.
Next, a configuration of the memory controller 20 will be described with reference to
The memory controller 20 will further be described with reference to
The ECC encoder 23 generates ECC data based on the encode ID and the write data given from the encoder 21. The ECC encoder 23 outputs the ECC data and the encode ID given from the cumulative value determination unit 22 to the page buffer 12. The ECC data and the encode ID are written in the ECC cell array 16. Particularly, in the memory block 14, the memory cell transistor MT connected to the bit line BL4313 is used to retain the encode ID.
The reading decoder 24 decodes the read data sensed and amplified by the sense amplifier 13 based on the encode ID. That is, when the encode ID read is information indicating that the non-inverting value is selected, the reading decoder 24 directly outputs the read data. When the encode ID read is information indicating that the inverted value is selected, the reading decoder 24 inverts each bit of the read data to output the data as the decoded data.
The write operation in the memory system having the above-described configuration will be described while taking note of the operations of the encoder 21 and the cumulative value determination unit 22. The operation of the encoder 21 will be described with reference to
Referring to
The operation of the cumulative value determination unit 22 will be described with reference to
When the received non-inverting value and inverted value are not the initial input data (NO in Step S22), the cumulative value determination unit 22 determines whether or not all the bits of the non-inverting value are “0”. When all the bits of the non-inverting value are “0” (YES in Step S26), the cumulative value determination unit 22 selects the inverted value to output the selection information as the encode ID to the encoder 21 and the ECC encoder 23 (Step S27). Accordingly, in this case, the inverted value is written in the regular cell array 15. The cumulative value determination unit 22 retains the adding result of each bit of the selected value and each digit of the cumulative value owned by itself as the new cumulative value (Step S28). That is, after Step S27, the cumulative value retained by the cumulative value determination unit 22 at that time is updated by the value to which the inverted value is further added. Then, the flow goes to Step S25.
When all the bits of the non-inverting value are not “0”, namely, when one of the bits is “1” (NO in Step S26), the cumulative value determination unit 22 determines whether or not all the bits of the non-inverting value are “1”. When all the bits of the non-inverting value are “1” (YES in Step S29), the cumulative value determination unit 22 selects the non-inverting value to output the selection information as the encode ID to the encoder 21 and the ECC encoder 23 (Step S30). Accordingly, in this case, the non-inverting value is written in the regular cell array 15. Then, the flow goes to the processing in Step S28. That is, after Step S30, the cumulative value retained by the cumulative value determination unit 22 is updated by the value to which the non-inverting value is further added. Then, the flow goes to Step S25.
When all the bits of the non-inverting value are not “1”, namely, when one of the bits is “0” (NO in Step S29), the flow goes to the processing in Step S31. In Step S31, the cumulative value determination unit 22 adds each digit of the cumulative value owned by itself and each bit of the non-inverting value and the inverted value received in Steps S20 and S21. The cumulative value determination unit 22 computes a difference between the maximum value and the minimum value of each digit in each adding result of the non-inverting value and the inverted value (Step S31). Then, the cumulative value determination unit 22 makes a determination of a magnitude relation of the difference obtained for each adding result of the non-inverting value and the inverted value obtained in Step S31 (Step S32). When the difference between the maximum value and the minimum value of the adding result for the non-inverting value is lower than the difference between the maximum value and the minimum value of the adding result for the inverted value (NO in Step S33), the cumulative value determination unit 22 selects the non-inverting value. The cumulative value determination unit 22 outputs the information indicating that the non-inverting value is selected as the encode ID to the encoder 21 and the ECC encoder 23 (Step S34). That is, the non-inverting value is written in the regular cell array 15.
On the contrary, the cumulative value determination unit 22 selects the inverted value, when the difference between the maximum value and the minimum value of the adding result for the non-inverting value is not lower than the difference between the maximum value and the minimum value of the adding result for the inverted value (YES in Step S33). The cumulative value determination unit 22 outputs the information indicating that the inverted value is selected as the encode ID to the encoder 21 and the ECC encoder 23 (Step S35). That is, the inverted value is written in the regular cell array 15.
After Steps S34 and S35, the cumulative value determination unit 22 goes to the processing in Step S28.
The processing of
First, the initial input data “1000” is input to the encoder. In this case, the non-inverting value and the inverted value become “1000” and “0111” respectively. Because the input data “1000” is the initial data written in the memory block 14 (YES in Step S22), the cumulative value determination unit 22 selects one of the non-inverting value and inverted value having more bits whose values are “1”, i.e., the inverted value “0111” (Step S23). The cumulative value becomes “0111”, and the encoder 21 and ECC encoder 23 are notified of “1” as the encode ID (Step S24).
Then, input data “1100” is input to the encoder. In this case, the non-inverting value and the inverted value become “1100” and “0011” respectively. The input data “1100” is not the initial data (NO in Step S22), all the bits of the non-inverting value of the input data “1100” are not “0” (NO in Step S26), and all the bits of the non-inverting value are not “1” (NO in Step S29). Therefore, the cumulative value determination unit 22 adds each digit of the cumulative value and each bit of the non-inverting value and inverted value (Step S31). Then, at this point, because the cumulative value is “0111”, the adding result of the cumulative value and the non-inverting value “1100” becomes “1211”. The adding result of the cumulative value and the inverted value “0011” becomes “0122”. In the adding result of the cumulative value and the non-inverting value, the maximum value of each digit becomes “2” while the minimum value becomes “1”. Therefore, the difference between the maximum value and the minimum value becomes “1”. In the adding result of the cumulative value and the inverted value, the maximum value of each digit becomes “2” while the minimum value becomes “0”. Therefore, the difference between the maximum value and the minimum value becomes “2”. Accordingly, the difference between the maximum value and the minimum value of “1” in the adding result for the non-inverting value is lower than the difference between the maximum value and the minimum value of “2” in the adding result for the inverted value (NO in Step S33). Accordingly, the cumulative value determination unit 22 selects the non-inverting value “1100” to set the encode ID to “0”. The cumulative value determination unit 22 retains the adding result “1211” for the inverted value as the new cumulative value (Step S28). In
The same processing as for the input number “2” is performed on the input numbers “3” and “4”.
Next, the input number “5” will be described. The fifth input data is “0000”. In this case, the non-inverting value and the inverted value become “0000” and “1111” respectively. Since the input data “0000” is not the initial data (NO in Step S22), and all the bits of the non-inverting value are “0” (YES in Step S26), the cumulative value determination unit 22 selects the inverted value “1111” (Step S27). At this point, because the cumulative value becomes “2322”, the adding result “3433” of the cumulative value “2322” and the inverted value “1111” becomes the new cumulative value (Step S28).
The same processing as the input number “2” is performed on the input numbers “6” and “7”.
Next, the input number “8” will be described. The eighth input data is “0011”. In this case, the non-inverting value and the inverted value become “0011” and “1100” respectively. The cumulative value determination unit 22 adds each digit of the cumulative value and each bit of the non-inverting value and the inverted value (Step S31). At this point, because the cumulative value becomes “4554”, the adding result of the cumulative value and the non-inverting value “0011” becomes “4565”. The adding result of the cumulative value and the inverted value “1100” becomes “5654”. In the adding result of the cumulative value and the non-inverting value, the maximum value of each bit becomes “6” and the minimum value becomes “5”. Therefore, the difference between the maximum value and the minimum value becomes “1”. In the adding result of the cumulative value and the inverted value, the maximum value of each digit becomes “6” and the minimum value becomes “5”, Therefore, the difference between the maximum value and the minimum value becomes “1”. Accordingly, the difference between the maximum value and the minimum value of “1” in the adding result for the non-inverting value is equal to the difference between the maximum value and the minimum value of “1” in the adding result for the inverted value (YES in Step S33). Accordingly, the cumulative value determination unit 22 selects the inverted value “1100” to set the encode ID to “1”. The cumulative value determination unit 22 retains the adding result “5654” for the inverted value as the new cumulative value (Step S28).
The same processing as for the input number “8” is performed on the input number “9”, and the same processing as for the input number “2” is performed on the input numbers “10” and “11”.
In the memory system according to the first embodiment of the invention, the following effects (1) and (2) are obtained.
(1) The operation reliability of the semiconductor storage device can be improved (Part 1).
In the configuration of the first embodiment, the memory controller 20 includes the cumulative value determination unit 22 which accumulates and retains the already-written data in each bit line. The cumulative value determination unit 22 determines whether the input data is directly written or the input data is written inverted according to the cumulative value. Accordingly, the operation reliability of the NAND type flash memory can be improved. The effect will be described below.
For example, in the NAND type flash memory disclosed in JP-A H7-334991 (KOKAI), a charge amount is averaged in the memory cell array as a whole. However, the charge amount is only averaged in the memory cell array as a whole, without considering any variation in the charge amount among the bit lines. Accordingly, the charges are possibly concentrated on one of the bit lines. When the charges are concentrated on one of the bit lines, a data error is easily created in the bit line. In the NAND type flash memory having the ECC function, it is necessary that ECC be performed to be able to aid the bit line in which the error is created most easily. Therefore, unfortunately, the ECC efficiency is lowered when the charges are concentrated on one of the bit lines.
On the contrary, in the configuration of the first embodiment, the data written in the memory cell transistor MT in the past is accumulated in each bit line. The non-inverting value and inverted value of the data to be newly written and the cumulative value are added, and the adding result in which the difference between the maximum value and the minimum value of each digit becomes the minimum is selected. That is, each digit of the adding result indicates the total number of the “1” data in each bit line. Accordingly, as the difference between the maximum value and the minimum value of each digit are increased in the adding result, the variation in charge amount is increased among the bit lines. Therefore, one of the non-inverting value and the inverted value is selected, such that the difference between the maximum value and the minimum value of each digit becomes smaller, namely, such that the variation in charge amount is decreased among the bit lines, and the selected value is written in the memory cell transistor MT.
Accordingly, the number of memory cell transistors MT which retain the “1” data is averaged among the bit lines. This enables the charges to be prevented from concentrating on one of the bit lines. Therefore, a probability of creating a data error is averaged among the bit lines to improve the ECC efficiency, which allows the operation reliability to be improved in the NAND type flash memory 10.
(2) The operation reliability of the semiconductor storage device can be improved (Part 2).
Recently, demands for large-capacity NAND type flash memories are increasing. At the same time, the reliability of write data and read data tends to be gradually lowered. This is attributed to the fact that, in the NAND type flash memory, a film thickness of the gate insulating film (tunnel oxide film) cannot be thinned lower than a predetermined value due to the data recording method using a tunnel current, whereas lateral sizes such as a gate length and a gate width of the memory cell transistor MT are reduced, whereby the recording data destruction caused by the GIDL phenomenon becomes pronounced.
Additionally, the number of memory cell transistors is increased in the NAND string because an area per bit is reduced. As a result, the current which can be passed through the NAND string is decreased, and a voltage change and current value which are required to read and sense the data are also decreased.
Increasing the number of memory cell transistors in the NAND string tends to lower the resistance against the read disturb (hereinafter abbreviated to RD) and program disturb (hereinafter abbreviated to PD). When the continuous memory cell transistors connected in series in the NAND string are turned on, the lowering of the resistance against RD and PD becomes more pronounced as the number of continuous memory cell transistors is increased. The following reasons can be cited for this.
In each memory cell transistor MT, there is a parasitic capacitance formed by a p-n junction between the source or drain region of the memory cell transistor and the well region in which the memory cell transistor is formed. When the plurality of memory cell transistors MT continuously connected in a column direction are turned on, a large parasitic capacitance is created in the NAND string. That is, the parasitic capacitances of the turned-on memory cell transistors are connected in parallel to form a very high parasitic capacitance. This very high parasitic capacitance has various negative effects on the NAND type flash memory.
For example, when the read voltage is applied to one of the word lines WL, the large amount of charge (current) supplied from the parasitic capacitance is accelerated to create hot electrons due to the gate voltage of the word line WL. The created hot electrons escape from the tunnel oxide film of the memory cell transistor connected to the word line WL, and destroy the data recorded in the memory cell transistor. This is the so-called GIDL phenomenon. Furthermore, when many continuous memory cell transistors are turned on, not only RD but also PD are easily generated.
However, in the configuration of the first embodiment, the input data is written while non-inverting inverted or inverted. In the input data, even if continuous pieces of “0” data exist in one of the bit lines, the “1” data which is the inverted value is written according to the determination result of the cumulative value determination unit 22, and the continuous pieces of “0” data can be prevented from existing on the same column. Accordingly, such continuous memory cell transistors can be prevented from being simultaneously turned on in the NAND string, which resolves the problem of RD and PD to improve the operation reliability of the NAND type flash memory.
In the first embodiment, the determination of the magnitude relation of the difference between the maximum value and the minimum value of each digit in the adding result of the cumulative value and the non-inverting value and inverted value is made in Steps S31 to S33. However, it is not always necessary that the determination processing be performed based on the maximum value and the minimum value of each digit in the adding result. For example, the determination processing may be performed using standard deviation.
As shown in
A semiconductor device according to a second embodiment of the invention will be described. In the second embodiment, “1” is subtracted from a value of each digit when all the digits of the cumulative value are at least “1” in the first embodiment. In the second embodiment, the configuration of the memory system 1 and the operation of the encoder 21 are similar to those of the first embodiment, therefore only the points different in the second embodiment will be described below.
As shown in
The processing of
Referring to
For the input number “3”, the cumulative value “0100” obtained in Step S50 and the non-inverting value “0110” or the inverted value “1001” are added. The adding result for the non-inverting value becomes “0210”, and the adding result for the inverted value becomes “1101”. Accordingly, the cumulative value determination unit 22 retains the adding result “1101” as the cumulative value. In this case, because the second digit of the cumulative value is “0”, the cumulative value determination unit 22 directly retains the adding result “1101” as the cumulative value without subtracting “1” from each digit in Step S50.
The same processing as above is applied for the input number “4” or more.
Thus, according to the memory system 1 of the second embodiment, the effect (3) is obtained in addition to the effects (1) and (2) described in the first embodiment.
(3) The size of the memory controller 20 can be reduced (Part 1).
In the case where the cumulative value is one in which pieces of data written in the past are simply accumulated, the maximum value is equal to the number of word lines in the NAND string in which the cumulative value is counted. For example, in the case of 32 word lines, the value of each digit becomes “32” at the maximum as the cumulative value. Therefore, each digit of the cumulative value becomes five-bit data in the binary number, and a five-stage register is required for each digit of the cumulative value.
On the other hand, in the configuration of the second embodiment, when the value of each digit of the cumulative value exceeds “1”, “1” is subtracted from all the digits. Therefore, the possibility that each digit of the cumulative value equals the available maximum value (“32” in the case of 32 word lines) is reduced, so that the number of registers can be decreased for each digit of the cumulative value to reduce the size of the memory controller 20.
An absolute value of the cumulative value has no meaning, and it is only necessary to know the difference in values between the bits of the cumulative value. This is because the difference in values between the digits of the cumulative value means the difference in the charge between the bit lines. Accordingly, the charge averaging among the bit lines is not affected even if “1” is subtracted from the value of each digit of the cumulative value like the embodiment.
In the second embodiment, the processing in Steps S31 to S33 can also be replaced by Steps S40 to S42 described in
A semiconductor device according to a third embodiment of the invention will be described. In the third embodiment, when an overflow is generated in one of the digits of the adding result in the second embodiment, the value of the digit generating the overflow is replaced by a quotient of “2” of the countable maximum value in the digit, and each of the other digits is replaced by a quotient of “2” of a retained value. In the third embodiment, the configuration of the memory system 1 and the operation of the encoder 21 are similar to those of the first and second embodiment, therefore only the points different in the third embodiment will be described below.
As shown in
When an overflow is generated (YES in Step S61), the cumulative value determination unit 22 sets the digit in which the overflow is generated to the available maximum value of the digit, and the cumulative value determination unit 22 sets the value in which all the digits are shifted rightward by one bit to the new adding result (Step S62). Assuming that “7” is the maximum value in a certain digit of the cumulative value, when the value of the digit becomes “8” to generate an overflow, the value of the digit is set to the maximum value “7” (“111” in the binary number). Then, all the digits are shifted rightward by one bit. That is, “111” is set to “011” (“3” in the decimal number). The resultant “3” is set as the adding result for the digit. In other words, the digit in which an overflow is generated is replaced by “3”, which is the quotient obtained by dividing the available maximum value “7” of the digit by “2”. Remainders of the division are rounded down. For other digits, the same processing is performed on the retained values.
The difference between the maximum value and the minimum value of each digit in the adding results for the non-inverting value and the inverted value is computed using the adding result obtained in Step S62 (Step S63), and the flow goes to Step S32. The processing from Step S32 is similar to that of the first and second embodiments.
Thus, according to the memory system 1 of the third embodiment, the effect (4) is obtained in addition to the effects (1) and (2) of the first embodiment and the effect (3) of the second embodiment.
(4) The size of the memory controller 20 can be reduced (Part 2).
Referring to
In this state, it is assumed that “1000” is further added. Because “1” is added to the retaining unit 43, although the adding result should be “3”+“1”=“4” under normal circumstances, the retaining unit 43 can count up to only “3”. That is, an overflow is generated, which results in incorrect addition.
On the other hand, in the third embodiment, a countermeasure to be taken in the case of an overflow is previously determined. Specifically, the quotient of “2” of the maximum value which can be retained by the retaining unit 43 is set as the value which should be retained when the overflow is generated. Accordingly, “1” which is the quotient obtained by dividing “3” by “2” is retained as shown in
Therefore, correct processing can be performed even if an overflow is generated. The processing of the third embodiment exerts a pronounced effect, in particular when the third embodiment is combined with the second embodiment. In the second embodiment, “1” is subtracted from each digit of the cumulative value to decrease the number of registers for retaining the cumulative value. In this case, the possible maximum value of each digit of the cumulative value is empirically determined. Therefore, when the number of registers is excessively decreased, an overflow is possibly generated in each digit of the cumulative value. Even in such cases, the countermeasure as in the third embodiment is taken against the overflow, which allows the processing to be correctly performed while the number of registers is effectively decreased.
In the NAND type flash memory 10, the pieces of data are sequentially written from the memory cell transistor MT (memory cell transistor MT connected to the word line WL0) located close to the select gate line SGS on the source side. Accordingly, the charge amounts among the bit line should be taken care of in the memory cell transistor MT close to the select gate line SGS. The overflow of the cumulative value is generated in the memory cell transistor MT in which the data is written late, i.e., in the memory cell transistor MT located close to the select gate line SGD on the drain side, and no overflow of the cumulative value is generated when data is written in the memory cell transistor MT located close to the select gate line SGS on the source side. Therefore, even if the difference in charge amount among the bit lines is not correctly represented in the cumulative value due to the generation of an overflow, the data write is already ended in the memory cell transistor MT located close to the select gate line SGS at that time, actually there is created no problem.
In the third embodiment, the processing in Steps S63 and S32 may be performed using the standard deviation described in the first embodiment of
A semiconductor device according to a fourth embodiment of the invention will be described. In the fourth embodiment, the encoder 21 is formed with a pseudo-random number generator in the first to third embodiments. Only the points different in the fourth embodiment will be described below.
Referring to
The first pseudo-random number generator 32 generates a pseudo-random number based on an upper bit of an address (hereinafter referred to as upper address) of the write data. That is, the first pseudo-random number generator 32 generates the pseudo-random number while the upper address is set as a scramble seed. The second pseudo-random number generator 33 generates the pseudo-random number based on a lower bit of the address (hereinafter referred to as low address) of the write data. That is, the second pseudo-random number generator 33 generates the pseudo-random number while the lower address is set as the scramble seed. The first and second pseudo-random number generators 32 and 33 have similar configurations, and the first and second pseudo-random number generators 32 and 33 generate the same pseudo-random number when the same scramble seed is used.
The XOR gate 34 performs an exclusive OR operation of the input data and the pseudo-random number generated by the first pseudo-random number generator 32, and the XOR gate 34 outputs the operation result as first scramble data. The XOR gate 35 performs an exclusive OR operation of the input data and the pseudo-random number generated by the second pseudo-random number generator 33, and the XOR gate 35 outputs the operation result as second scramble data.
The selector 36 outputs one of the first and second pieces of scramble data as the write data according to the encode ID. Specifically, the selector 36 selects the first scramble data when the encode ID is “0”, and the selector 36 selects the second scramble data when the encode ID is “1”.
Referring to
The selector 60 selects one of the upper address and the lower address according to the encode ID. Specifically, the selector 60 selects the upper address when the encode ID is “0”, and the selector 36 selects the lower address when the encode ID is “1”. The pseudo-random number generator 61 has a configuration similar to those of first and second pseudo-random number generators 32 and 33 in the encoder 21, and the pseudo-random number generator 61 generates a pseudo-random number while setting one of the upper address and the lower address selected by the selector 60 as the scramble seed. The XOR gate 62 performs an exclusive OR operation of the read data sensed and amplified by the sense amplifier 13 and the pseudo-random number generated by the pseudo-random number generator 61. The operation result in the XOR gate 62 becomes the decoded data.
The operation of the encoder 21 of the fourth embodiment is one in which the non-inverting value and the inverted value are replaced by the first and second pieces of scramble data respectively in the flowchart of the first embodiment of
The operation of the cumulative value determination unit 22 of the fourth embodiment is one in which the non-inverting value and the inverted value are replaced by the first and second pieces of scramble data respectively in the flowcharts of the first to third embodiments of
The input data is scrambled according to different rules by the first and second pseudo-random number generators 32 and 33. The generated first and second pieces of scramble data are written as the write data in the NAND type flash memory 10.
The cumulative value determination unit 22 accumulates the first and second pieces of scramble data written in the NAND type flash memory 10. Each digit of the cumulative value corresponds to each bit line in the NAND type flash memory 10. The cumulative value determination unit 22 adds each digit of the cumulative value and each bit of the first scramble data, and the cumulative value determination unit 22 adds each digit of the cumulative value and each bit of the second scramble data. The cumulative value determination unit 22 selects the difference between the maximum value and the minimum value of each digit of the adding result in the first and second pieces of scramble data, or the cumulative value determination unit 22 selects one of the maximum value and the minimum value having the smaller standard deviation.
According to the memory system 1 of the fourth embodiment, the effect (5) is obtained in addition to the effects (1) to (4) described in the first to third embodiments.
(5) The operation reliability of the semiconductor storage device can be improved (Part 3).
In the memory system 1 of the fourth embodiment, the encoder 21 includes the linear feedback shift register, and the linear feedback shift register includes the first and second pseudo-random number generators 32 and 33 and the XOR gate 34 and 35. One of the first and second pieces of scramble data, obtained by performing the scramble on the input data using the linear feedback shift register, is written in the memory cell transistor MT.
That is, the data written in the NAND type flash memory 10 is randomized in the bit line direction based on which the first and second pieces of scramble data is selected. The scramble is performed using a pseudo-random number to randomize the data in the page. In other words, the data is also randomized in the word line direction.
Accordingly, the same pieces of data can be prevented from continuously existing not only in the bit line direction but also in the word line direction. As described in the effect (2) of the first embodiment, the problem of RD or PD can be reduced to improve the operation reliability of the NAND type flash memory.
In the fourth embodiment, the encoder 21 includes the two pseudo-random number generators. Alternatively, the encoder 21 may include the one pseudo-random number generator. In this case, the one pseudo-random number generator sets the upper address and the lower address as the scramble seed to generate the two pseudo-random numbers, and the two pseudo-random numbers are retained by the registers. Then, the exclusive OR operation of the pseudo-random number retained by the register and the input data is performed.
The encoder 21 may generate at least three pieces of scramble data. In this case, the one pseudo-random number generator may generate at least three random numbers, or at least three pseudo-random number generators may be provided.
A semiconductor device according to a fifth embodiment of the invention will be described. The fifth embodiment relates to the case in which each of the memory cell transistors MT retains four-value (2-bit) data in the first to third embodiments. Only the points different in the fifth embodiment will be described below.
A threshold of the memory cell transistor MT of the fifth embodiment will be described with reference to
As shown in
A method for writing four-value data in the memory cell MC will briefly be described with reference to
At the start of the write, the memory cells are in the erased state. That is, the memory cells have the negative threshold voltage Vth, and the memory cells retain the “11” data. In this state, the first write is performed. When the write data (first write data) is “1”, electrons are not injected into the floating gate, the threshold voltage Vth is not changed. When the first write data is “0”, electrons are injected into the floating gate, the threshold voltage Vth is changed toward the positive direction, and the threshold voltage Vth becomes Vth1<Vth<Vth2. That is, the memory cell has substantially the same threshold as that obtained in retaining the “00” data.
Then, the second write is performed. First the case in which the first write data is “1” will be described. When the second write data is “1”, in the second write, electrons are not injected into the floating gate and the threshold voltage Vth is kept at a negative value. As a result, the “11” data is written in the memory cell. When the second write data is “0”, electrons are injected into the floating gate, which changes the threshold voltage Vth to the positive direction. The threshold voltage Vth is 0V<Vth<Vth1. That is, the “10” data is written in the memory cell.
Next, the case in which the first write data is “0” will be described. When the second write data is “0”, because electrons are not injected into the floating gate in the second write, the result of the first write is maintained. The threshold voltage Vth is Vth1<Vth<Vth2, and the “00” data is written in the memory cell. When the second write data is “1”, electrons are further injected into the floating gate, which further changes the threshold voltage Vth to the positive direction. The threshold voltage Vth becomes Vth2<Vth<Vth3. That is, the “01” data is written in the memory cell.
The configuration of the encoder 21 included in the memory controller 20 of the fifth embodiment will be described with reference to
The memory 39 includes an encoding table 70. An example of the encoding table will be described with reference to
The first encoding unit 37 encodes the non-inverting value to generate first encoding data according to the encoding table 70 of the memory 39, and the first encoding unit 37 outputs the first encoding data to the cumulative value determination unit 22. That is, the first encoding unit 37 outputs “0”, “1”, “2”, and “3” as the first encoding data when the non-inverting value are “01”, “00”, “10”, and “11” respectively.
The second encoding unit 38 encodes the inverted value to generate second encoding data according to the encoding table 70 of the memory 39, and the second encoding unit 38 outputs the second encoding data to the cumulative value determination unit 22. That is, the second encoding unit 38 outputs “0”, “1”, “2”, and “3” as the second encoding data when the inverted value are “01”, “00”, “10”, and “11” respectively.
Other configurations are similar to those of the first to third embodiments, therefore their descriptions are omitted.
The operation of the encoder 21 having the above-described configuration will be described with reference to
Referring to
Basically, the operation of the cumulative value determination unit 22 is one in which the non-inverting value and the inverted value are replaced by the first and second pieces of encoding data in the flowcharts of
Referring to
When the first and second pieces of encoding data are not the initial input data (NO in Step S92), the cumulative value determination unit 22 determines whether or not all the non-inverting values are equal to one another and the upper bit of the non-inverting value is “0”. The determination whether or not all the non-inverting values are equal to one another can determine whether or not all the bits of the first encoding data have the same value. The determination of whether or not the upper bit is “0” may be made by inputting the non-inverting value into the cumulative value determination unit 22 aside from the first and second pieces of encoding data, or the cumulative value determination unit 22 may make the determination of whether or not the upper bit is “0” from the first encoding data based on the correlation shown in
When conditions of Step S96 are not satisfied (NO in Step S96), the cumulative value determination unit 22 determines whether or not all the non-inverting values are equal to one another and the upper bit of the non-inverting value is “1”. When all the non-inverting values are equal to one another and the upper bit of the non-inverting value is “1” (YES in Step S99), the cumulative value determination unit 22 selects the first piece of encoding data to output the selection information as the encode ID to the encoder 21 and the ECC encoder 23 (Step S100). Accordingly, in this case, the non-inverting value is written in the regular cell array 15. Then, the flow goes to processing in Step S98. That is, after Step S100, the cumulative value retained by the cumulative value determination unit 22 at that time is updated by the value to which the first encoding data is further added. Then, the flow goes to Step S95.
When the conditions of Step S99 are not satisfied, namely, when all the non-inverting values are not equal to one another (NO in Step S99), the flow goes to the processing in Step S101. In Step S101, the cumulative value determination unit 22 adds each digit of the cumulative value owned by itself and each bit of the first and second pieces of encoding data received in Steps S90 and S91. The cumulative value determination unit 22 calculates the difference between the maximum value and the minimum value of each digit in each adding result of the first and second pieces of encoding data (Step S101). Then, the cumulative value determination unit 22 makes a determination of the magnitude relation of the difference obtained for each adding result of the first and second pieces of encoding data obtained in Step S101 (Step S102). When the difference between the maximum value and the minimum value of the adding result for the first encoding data is lower than the difference between the maximum value and the minimum value of the adding result for the second encoding data (NO in Step S103), the cumulative value determination unit 22 selects the first encoding data. The cumulative value determination unit 22 outputs the information indicating that the first encoding data is selected as the encode ID to the encoder 21 and the ECC encoder 23 (Step S104). That is, the non-inverting value is written in the regular cell array 15.
On the contrary, the cumulative value determination unit 22 selects the second encoding data, when the difference between the maximum value and the minimum value of the adding result for the first encoding data is not lower than the difference between the maximum value and the minimum value of the adding result for the second encoding data (YES in Step S103). The cumulative value determination unit 22 outputs the information indicating that the second encoding data is selected as the encode ID to the encoder 21 and the ECC encoder 23 (Step S105). That is, the inverted value is written in the regular cell array 15.
After Steps S104 and S105, the cumulative value determination unit 22 goes to the processing in Step 98.
A specific example of the processing will be described below with reference to
First, initial pieces of input data “01”, “11”, “10”, and “11” are input to the encoder 21. In this case, the inverted values become “10”, “00”, “01”, and “00”. The first encoding data becomes “0323” and the second encoding data becomes “2101” according to the rule of
Then, pieces of input data “00”, “01”, “10”, and “00” are input to the encoder 21. In this case, the inverted values become “11”, “10”, “10”, and “11”. The first and second pieces of encoding data become “1021” and “3203” respectively. The pieces of input data “00”, “01”, “10”, and “00” are not the initial data (NO in Step S92), and all the non-inverting values are not equal to one another (NO in Step S96, NO in Step S99). Therefore, the cumulative value determination unit 22 adds each digit of the cumulative value and each bit of the first and second pieces of encoding data (Step S101). Then, at this point, because the cumulative value is “0323”, the adding result of the cumulative value and the first encoding data becomes “1344”. The adding result of the cumulative value and the second encoding data becomes “3526”. The difference between the maximum value and the minimum value “3” of each digit of the adding result of the cumulative value and the first encoding data is lower than the difference between the maximum value and the minimum value “4” of each digit of the adding result of the cumulative value and the second encoding data (NO in Step S103). Accordingly, the cumulative value determination unit 22 selects the first encoding data “1344” to set the encode ID to “0”. The cumulative value determination unit 22 retains the adding result “1344” for the first encoding data as the new cumulative value (Step S98). When “0” is set as the encode ID, the selector 31 of the encoder 21 selects the non-inverting value to output the non-inverting values “00”, “01”, “10”, and “00” as the write data.
The processing similar to that of the input number “2” is performed on the input numbers “3” and “4”.
Next, the input number “5” will be described. Fifth pieces of input data are “11”, “11”, “11”, and “11”. In this case, the inverted values become “00”, “00”, “00”, and “00” respectively. The first and second pieces of encoding data become “3333” and “1111” respectively. The pieces of input data “11”, “11”, “11”, and “11” are not the initial data (NO in Step S92), the non-inverting values are equal to one another, and the upper bit of the non-inverting value is “1” (NO in Step S96 and YES in Step S99). Accordingly, the cumulative value determination unit 22 selects the first encoding data “3333” (Step S100). Then, at this point, because the cumulative value is “7665”, the adding results “10”, “9”, “9”, and “8” of the cumulative value “7665” and the first encoding data “3333” becomes the new cumulative value (Step S98).
Next, the input number “6” will be described. Sixth pieces of input data are “00”, “00”, “00”, and “00”. In this case, the inverted values become “11”, “11”, “11”, and “11” respectively. The first and second pieces of encoding data become “1111” and “3333” respectively. The pieces of input data “00”, “00”, “00”, and “00” are not the initial data (NO in Step S92), the non-inverting values are equal to one another, and the upper bit of the non-inverting value is “0” (YES in Step S96). Accordingly, the cumulative value determination unit 22 selects the second encoding data “3333” to set the encode ID to “1”. Then, the cumulative value determination unit 22 retains the adding results “13”, “12”, “12”, and “11” as the new cumulative value for the second encoding data (Step S97). When the encode ID is set to “1”, the selector 31 of the encoder 21 selects the inverted value to output the inverted values “11”, “11”, “11”, and “11” as the write data.
Thus, in the memory system of the fifth embodiment, the first and second encoding units 37 and 38 encode the multi-bit data according to the charge amount of the multi-bit data. More specifically, the encoding value is increased in the ascending order of the charge amount injected into the floating gate when the multi-bit data is written (see
The fifth embodiment can be combined with the fourth embodiment. Referring to
A semiconductor device according to a sixth embodiment of the invention will be described. The sixth embodiment relates to another coding method of the fifth embodiment. Only the different points in the sixth embodiment will be described below.
The operation of the encoder 21 will be described with reference to
Referring to
Thus, in the memory system of the sixth embodiment, the effects (1) to (4) are obtained like the fifth embodiment. The following effect (6) is additionally obtained.
(6) The size of the memory controller 20 can be reduced (Part 3).
In the configuration of the sixth embodiment, the first and second encoding units 37 and 38 output the upper bits of the non-inverting value and the inverted value as the first and second pieces of encoding data. Accordingly, when the input data is the two-bit data, the first and second pieces of encoding data have the available values of “0” and “1”. Therefore, the cumulative value can be decreased compared with the fifth embodiment. Specifically, in the case of 32 word lines in the fifth embodiment, the available maximum value of each digit of the cumulative value is (“3”×32)=“96”. On the other hand, in the sixth embodiment, the available maximum value of each digit of the cumulative value is (“1”×32)=“32”. Therefore, the size of the register retaining the cumulative value can be reduced to achieve downsizing of the memory controller 20.
In the above description, the encoder 21 retains the encoding table 70 in the memory 39. However, the encoding table 70 is not required as long as the first and second encoding units 37 and 38 can take out the upper bits of the non-inverting value and the inverted value. Similarly to the fifth embodiment, in the sixth embodiment, the first and second pieces of scramble data may be input to the first and second encoding units 37 and 38 instead of the non-inverting value and the inverted value.
A semiconductor device according to a seventh embodiment of the invention will be described. The seventh embodiment relates to a configuration in which the page data is randomized in the word line direction while the charge amount among the bit lines is not averaged by the cumulative value.
Referring to
The encoder 21 of the seventh embodiment as the configuration in which the second pseudo-random number generator 33, the XOR gate 35, and the selector 36 are eliminated in the configuration of the fourth embodiment of
The ECC encoder 23 generates the ECC data based on the scramble data output from the XOR gate 34.
The reading decoder 24 has the configuration in which the selector 60 is eliminated from the configuration of the fourth embodiment of
According to the memory system 1 of the seventh embodiment, the following effect (7) is obtained.
(7) The effects (1) and (5) are obtained by the simple configuration.
In the configuration of the seventh embodiment, the input data is scrambled by the pseudo-random number generated by the pseudo-random number generator 32, and the scrambled data is written in the NAND type flash memory 10. That is, the write data is randomized in the word line direction. Accordingly, the same pieces of data can be prevented from continuously existing along the word line direction, and the effect (5) described in the fourth embodiment is obtained.
The input data is scrambled by the pseudo-random number, which averages the charge amount among the bit lines. Accordingly, the effect (1) described in the first embodiment is obtained. Obviously, for averaging the charge amount, the method of the first embodiment for selecting the write data based on the cumulative value of the charge amount of the bit line is superior to the method of the seventh embodiment. On the other hand, the cumulative value determination unit 22 and the selector 31 of the encoder 21 are not required in the seventh embodiment. Accordingly, in the case where a simple circuit configuration is strongly demanded, it is desirable to adopt the configuration of the seventh embodiment. In the seventh embodiment, because the encode ID is not used, the memory cell transistor MT connected to the bit line BL4313 in the ECC cell array 16 can be used to retain the ECC data.
In the semiconductor devices of the first to sixth embodiment, the cumulative value determination unit 22 of the memory controller 20 retains the cumulative value in which each digit is the accumulating result of the charge amount in each bit line. As the value of each digit of the cumulative value is increased, the charge amount in the corresponding bit line is decreased. As the value of each digit is decreased, the charge amount is increased. The cumulative value determination unit 22 selects the write data such that the difference among the digits of the cumulative value is decreased, i.e., averaged. As a result, the charge amounts are averaged among the bit lines, and the problem of PD or RD can be solved to improve the operation reliability of the semiconductor device. In the seventh embodiment, the same effect is also obtained by randomizing the write data in the word line direction.
Both the case in which the encoder 21 outputs the non-inverting value and the inverted value and the case in which the first and second pieces of scramble data are output are described in the above embodiments. Alternatively, both the cases may be combined.
Referring to
In this case, the cumulative value determination unit 22 adds each bit of the cumulative value and each bit of the inverted value, each bit of the non-inverting value, and each bit of the first and second pieces of scramble data. The cumulative value determination unit 22 selects one of the inverted value, the non-inverting value, and the first and second pieces of scramble data in which the difference between the maximum value and the minimum of each digit of the adding result becomes the minimum. Thus, the four pieces of data may be set to the write candidate to select the optimum data.
Accordingly, the encoder 21 in the first to sixth embodiments can be represented by a configuration of
In this case, the cumulative value determination unit 22 adds each bit of the cumulative value and each bit of each of the 1st to 2 k-th write data candidates. The cumulative value determination unit 22 selects one of the 1st to 2 k-th write data candidates in which the difference between the maximum value and the minimum of each digit of the adding result becomes the minimum.
The above embodiments can be applied to various NAND type flash memory devices. The case in which the first to seventh embodiments are applied to a memory card will be described below.
Referring to
The memory card 110 is connected to the memory controller 20 through a bus interface 122. The memory card 110 includes the NAND type flash memory 10 described in the first embodiment and a plurality of signal pins 111. The signal pin 111 can electrically be connected to the memory controller 20. The electric power and various signals are given from the memory controller 20 through the signal pins 111.
In the above configuration, the bus interfaces 121 and 122 have eight-bit data lines. In the case where one-page (4096 bytes) data is written, the one-page data is transferred 4096 times while divided into each eight bits. The line buffer 25 of the memory controller 20 is used to perform buffering of the transfer data. The one-page data buffered by the line buffer 25 is input to the encoder 21. Thus, the processing described in the first to seventh embodiments may be performed outside the memory card 110.
In this case, the processing described in the first to seventh embodiments and the generation of the ECC data are performed inside the memory card 130. The write data and the ECC data is transferred to the page buffer 12 of the NAND type flash memory 10 though an internal bus connecting the memory controller 20 and the NAND type flash memory 10. Similarly to the case of
In the above embodiments, the cumulative value is described as the cumulative result of the “1” data. Alternatively, the cumulative value may be the cumulative result of the “0” data. That is, the absolute value of the cumulative value has no meaning, and it is only necessary to know the difference in charge amount between the digits of the cumulative value. In the fifth embodiment, each memory cell retains the two-bit data (four values). Alternatively, each memory cell may retain the data having at least three bits (eight values). In such cases, the data can be dealt with by the same methods as that of
Additional advantages and modifications will readily occur to those skilled in the art. Therefore, the invention in its broader aspects is not limited to the specific details and representative embodiments shown and described herein. Accordingly, various modifications may be made without departing from the spirit or scope of the general inventive concept as defined by the appended claims and their equivalents.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2007-050390 | Feb 2007 | JP | national |
| Number | Name | Date | Kind |
|---|---|---|---|
| 6222762 | Guterman et al. | Apr 2001 | B1 |
| 6279133 | Vafai et al. | Aug 2001 | B1 |
| 6728825 | Norman | Apr 2004 | B1 |
| 7126850 | Tatsukawa et al. | Oct 2006 | B2 |
| 7330373 | Lee et al. | Feb 2008 | B2 |
| 7359249 | Urabe et al. | Apr 2008 | B2 |
| 20020159285 | Morley et al. | Oct 2002 | A1 |
| 20060181934 | Shappir et al. | Aug 2006 | A1 |
| 20070130496 | Kanno | Jun 2007 | A1 |
| 20070174740 | Kanno | Jul 2007 | A1 |
| Number | Date | Country |
|---|---|---|
| 6076586 | Mar 1994 | JP |
| 7-334991 | Dec 1995 | JP |
| 8-50562 | Feb 1996 | JP |
| 2004-046985 | Feb 2004 | JP |
| 2006-139884 | Jun 2006 | JP |
| 2006-228407 | Aug 2006 | JP |
| Number | Date | Country | |
|---|---|---|---|
| 20080205145 A1 | Aug 2008 | US |
