Patent Application
20080232181
This application is based upon and claims the benefit of priority from the prior Japanese Patent Applications No. 2007-74783, filed on Mar. 22, 2007, the entire contents of which are incorporated herein by reference.
1. Field of the Invention
The present invention relates to a semiconductor memory device. For example, the invention relates to a semiconductor memory device which has FBCs (Floating Body Cells).
2. Related Art
In recent years, there is an FBC memory device as a semiconductor memory device expected as a memory replacing a 1T(Transistor)-1C (Capacitor) type DRAM. The FBC memory device has an FET (Field Effect Transistor) having a floating body (hereinafter, also “body”) formed on an SOI (Silicon On Insulator) substrate, and stores data “1” or data “0” depending on the number of majority carriers accumulated in this body.
As test items of the FBC memory, there are a static characteristic of a memory cell, a writing characteristic, a data retention characteristic, and a threshold voltage distribution. A test result of a memory cell formed in various process conditions is used to determine an optimum process condition and to analyze a defect cause. Many memory cells within a memory cell array need to be tested. To test the characteristics of these many memory cells, wirings for directly accessing the memory cells from an external pad via a DQ line are provided.
In the case of measuring a threshold voltage of a memory cell, for example, a change of a current flowing from the external pad to the memory cell is measured by controlling a gate voltage (word line voltage) of the memory cell. That is, a testing device directly applies a voltage to the memory cell through an external pad, without using a sense amplifier or a DQ buffer, thereby directly measuring the flowing current. The testing device changes the word line voltage from a low voltage to a high voltage. Further, the testing device uses as a threshold voltage the word line voltage when a predetermined current flows from the external pad to the memory cell. By repeatedly performing this to all memory cells, a distribution of the threshold voltage can be obtained.
However, usually because there is a variation in the threshold voltage of a memory cell, a current flowing from the external pad to the memory cell at a certain word line voltage is different for each bit. When the word line voltage is changed, a current flowing from the external pad to the memory cell also changes. Therefore, it has been difficult to simultaneously test all memory cells connected to a certain word line. Consequently, conventionally, the test needs to be executed to each bit. Accordingly, a long measuring time is necessary to test all the memory cells. Further, when the number of memory cells within the memory cell array increases, the test time becomes longer.
A semiconductor memory device according to an embodiment of the present invention comprises a memory cell array having memory cells arrayed two-dimensionally; word lines connected to the memory cells of rows of the memory cell array; bit lines connected to the memory cells of columns of the memory cell array; sense amplifiers connected to the bit lines, and detecting data stored in the memory cells; a test pad passing a predetermined reference current from a power source, and transmitting a reference voltage based on the reference current; and test circuits connected between the power source and the test pad and intervening between the power source and the bit lines, the test circuits passing test currents according to the reference voltage via the bit lines.
Embodiments of the present invention will be explained below in detail with reference to the accompanying drawings. Note that the invention is not limited thereto.
The memory cell array MCA includes plural memory cells arrayed two-dimensionally in a matrix shape. Word lines WLi (i is an integer) (hereinafter, also WLs) extend in a row direction, and are connected to gates of memory cells MCs of the rows of the memory cell array MCA. The integer i is 0 to 255, for example, and only WL0 and WL1 are shown in
Each sense amplifier S/A is provided corresponding to each bit line pair BL, bBL. As explained above, the present embodiment employs what is called a two-cell/bit system (also called a twin cell). This system stores data of opposite logics in two memories MCs forming a pair, thereby storing one-bit data. The sense amplifier S/A uses one of the data of the opposite logic as a basis, and detects the other data. However, the present invention can be also applied to a device having a one-cell/bit system or an open-bit line structure, as well as a device having the two-cell/bit system.
The sense amplifier S/A includes a pair of sense nodes SN, bSN. The sense node pair SN, bSN is connected to a corresponding bit line pair BL, bBL via transfer gates N5, N6. The transfer gates N5 and N6 are on/off controlled by signals FAIT0 and FAIT1, respectively. To test the memory cell MC connected to the bit line BL, the transfer gate N5 is set to the on state, and the sense node SN is connected to the bit line BL. In this case, the transfer gate N6 is in the off state. To test the memory cell MC connected to the bit line bBL, the transfer gate N6 is set to the on state, and the sense node bSN is connected to the bit line bBL. In this case, the transfer gate N5 is in the off state.
The sense amplifier S/A includes cross-couple type dynamic latch circuits (hereinafter, also “latch circuits”) LC1 and LC2. The latch circuit LC1 includes two p-type transistors P9 and P10 connected in series between the sense nodes SN and bSN. A gate of the transistor P9 is connected to the sense node bSN, and a gate of the transistor P10 is connected to the sense node SN. That is, the gates of the transistors P9 and P10 are cross-connected to the sense nodes SN and bSN. The latch circuit LC2 includes two n-type transistors N3 and N4 connected in series between the sense nodes SN and bSN. A gate of the transistor N3 is connected to the sense node bSN, and a gate of the transistor N4 is connected to the sense node SN. That is, the gates of the transistors N3 and N4 are also cross-connected to the sense nodes SN and bSN. The latch circuits LC1 and LC2 are driven by activating signals SAP and bSAN.
The activation means to turn on or drive an element or a circuit, and inactivation means to turn off or stop an element or a circuit. Therefore, it should be noted that a HIGH (high-potential level) signal can be an activation signal, and a LOW (low-potential level) signal can be an activation signal. For example, an NMOS transistor is activated by setting the gate to HIGH. On the other hand, a PMOS transistor is activated by setting the gate to LOW. The signal SAP activates the latch circuit LC1 when the signal SAP is at the high-potential level. The signal bSAN activates the latch circuit LC2 when the signal bSAN is at the low-potential level.
The n-type transistor N1 is connected between the sense node SN and the DQ line DQ0, and the n-type transistor N2 is connected between the sense node bSN and the DQ line bDQ0. The transistors N1, N2 are controlled by a column selection line CSL. The DQ line pair DQ0, bDQ0 is connected to a DQ buffer (not shown). The DQ buffer is connected to an I/O pad. At the time of reading data to an external of the memory device, the DQ buffer temporarily stores data from the memory cell MC to output the data to the outside. At the time of writing data from the external of the memory device, the DQ buffer temporarily stores data from the outside to transmit the data to the sense amplifier S/A. Accordingly, the column selection line CSL is activated at the time of reading data to the outside of the memory device or at the time of writing data from the outside of the memory device, thereby making it possible to connect the sense nodes SN and bSN to the DQ buffer.
In the operation of data reading to the outside of the memory device (normal reading operation), the sense amplifier S/A detects data stored in the memory cell MC, and the latch circuits LC1, LC2 latch this data to the sense node pair SN, bSN. The data latched into the sense nodes SN, bSN is temporarily stored in the DQ buffer via the DQ line, and is output from the I/O pad to the outside. In the operation of data writing from the outside of the memory device (normal writing operation), the DQ buffer temporarily stores data input from the I/O pad. This data is transmitted to the sense amplifier S/A via the DQ line. The latch circuits LC1, LC2 latch this data to the sense node pair SN, bSN. The data latched into the sense nodes SN, bSN is written into the memory cell MC via the bit line pair BL, bBL.
The test circuit includes a current mirror circuit CM including p-type transistors P1 to P4, selection transistors P5, P6, and transfer gates P7, P8. The transistors P1 and P3 are switching elements that are on/off controlled by a signal bLOADON. The transistors P1 and P3 become in the on state when a current is passed from the current mirror circuit CM at the test time. The on resistance of the transistors P1 and P3 is negligibly smaller than the on resistance of the transistors P2 and P4.
The transistor P2 as a first transistor and the transistor P4 as a second transistor are present between power sources VDDs via the transistors P1 and P3 and a test pad PAD, or are present between the power sources VDDs via the transistors P1 and P3 and the sense node SN or bSN. A connection relationship between the transistors P2 and P4 changes depending on which one of memory cells connected to the bit lines BL and bBL is to be tested. The transistors P2 and P4 are structured to have a predetermined mirror ratio. The transistor P4 uses a current flowing through the transistor P2 as a reference, and passes a current corresponding to this reference current. Alternatively, the transistor P2 uses a current flowing through the transistor P4 as a reference, and passes a current corresponding to this reference current.
The selection transistor P5 as a fourth transistor is connected between the transistor P4 and the test pad PAD, and is controlled by a test signal TS0. The selection transistor P6 as a third transistor is connected between the transistor P2 and the test pad PAD, and is controlled by a test signal TS1. A node between the selection transistor P5 and the transistor P4 is connected to the sense node SN via the transfer gate P7. A node between the selection transistor P6 and the transistor P2 is connected to the sense node bSN via the transfer gate P8.
When a node between the sense node SN and the transistor P4 is called a first node and also when a node between the sense node bSN and the transistor P2 is called a second node, the selection transistors P5 and P6 are connected in series between the first node and the second node. A third node between the selection transistors P5 and P6 is connected to the test pad PAD.
The selection transistor P5 is turned on at the time of testing the memory cell MC connected to the bit line bBL, thereby connecting a source or a drain of the selection transistor P4 and the sense node SN to the test pad PAD. The selection transistor P6 is turned on at the time of testing the memory cell MC connected to the bit line BL, thereby connecting a source or a drain of the selection transistor P2 and the sense node bSN to the test pad PAD.
The transfer gate P7 including a p-type transistor is connected between the current mirror circuit CM and the sense node SN, and the transfer gate P8 including a p-type transistor is connected between the current mirror circuit CM and the sense node bSN. The transfer gates P7 and P8 are controlled by a signal TFAIT.
A first test for testing the memory cells MCs connected to the bit lines BLj and a second test for testing the memory cells MCs connected to the bit lines bBLj are explained next. It is assumed that the transistors P1, P3, P7, and P8 are all in the on state during the execution of the first and the second tests.
In the first test, the selection transistor P5 becomes in the off state, and the selection transistor P6 becomes in the on state. In this case, one of the drain and the source of the transistor P2 is connected to the power source VDD via the transistor P1. The other of the drain and the source of the transistor P2 is connected to the test pad PAD via the transistor P6. Accordingly, the transistor P2 makes a first current path indicated by an arrowhead CP1 conductive, and passes the current (reference current) from the power source VDD to the test pad PAD via the transistor P3. One of the drain and the source of the transistor P4 is connected to the power source VDD. The other of the drain and the source of the transistor P4 is connected to the sense node SN via the transfer gate P7. Accordingly, the transistor P4 makes a second current path indicated by an arrowhead CP2 conductive, and passes the test current to the memory cell concerned.
In the second test, the selection transistor P5 becomes in the on state, and the selection transistor P6 becomes in the off state. In this case, one of the drain and the source of the transistor P2 is connected to the power source VDD via the transistor P1. The other of the drain and the source of the transistor P2 is connected to the sense node bSN via the transfer gate P8. Accordingly, the transistor P2 makes a second current path indicated by an arrowhead CP12 conductive, and passes the test current to the memory cell concerned. One of the drain and the source of the transistor P4 is connected to the power source VDD via the transistor P3. The other of the drain and the source of the transistor P4 is connected to the test pad PAD via the transistor gate P5. Accordingly, the transistor P4 makes a first current path indicated by an arrowhead CP11 conductive, and passes the current (reference current) from the power source VDD to the test pad PAD. As explained above, the roles of the transistors P2 and P4 are replaced with each other between the first test and the second test.
As explained above, the current mirror circuit CM includes the first current path CP1 or CP11 and the second current path CP2 or CP12. In the first test for testing the memory cell MC connected to the bit line BL, the first and the second current paths become the paths indicated by the arrowheads CP1 and CP2 in
The test pad PAD is connected to the current mirror circuit CM within the test circuit TC. The test pad PAD is commonly used for plural test circuits TCs provided in plural columns within the memory cell array MCA. For example, one test pad PAD is provided within the memory cell array MCA. The test pad PAD is connected to the first path CP1 or CP11 for passing the reference current. The reference current is a constant current becoming the reference at the time of detecting data of the memory cell MC, and does not change. The reference current can be equal in the test circuit TC of each column. That is, the test pad PAD can supply a constant current, and does not need to differentiate the current supplied to each test circuit TC. Therefore, the test pad PAD can be made common to the plural test circuits TCs.
One of methods for writing data into the memory cell MC is explained below. To write data “1” into the memory cell MC, the memory cell MC is operated in a saturated state. For example, the word line WL is biased at 1.5 V, and the bit line BL is biased at 1.5 V. The source is VSL (for example, ground GND (0V)). Accordingly, impact ionization occurs near the drain, and a large amount of electron-hole pairs occur. Electrons generated by impact ionization flow to the drain, and holes are stored in the body having low potential. When the current flowing at the time of the occurrence of the holes by impact ionization is balanced with the forward current in a pn junction between body sources, the body voltage reaches an equilibrium state. This body voltage is about 0.7 V.
At the time of writing the data “0”, the bit line BL is decreased to a negative voltage. For example, the potential of the bit line BL is dropped to −1.5 V. By this operation, the pn junction between the body 50 and the drain 40 is greatly biased to a forward direction. The holes accumulated in the body 50 are discharged to the drain 40, and the data “0” is stored in the memory cell MC.
One of methods for reading data from the memory cell MC is explained below. In the data reading operation, the word line WL is activated simultaneously with the data writing. However, the bit line BL is set lower than that at the time of writing data “1”. For example, the word line WL is set to 1.5V, and the bit line BL is set to 0.2 V. The memory cell MC is operated in a linear region. The memory cell MC for storing the data “0” is different from the memory cell MC for storing the data “1” in the threshold voltages of the memory cells MCs due to the difference in the numbers of holes accumulated in the body 50. By detecting a difference between the threshold voltages, the data “1” and the data “0” are discriminated. A reason why the bit line BL is set to the low voltage at the reading time is that when the memory cell MC is biased to a saturated state by increasing the voltage of the bit line BL, there is a risk that the data “0” is changed to the data “1” by the impact ionization at the time of reading the data “0”.
The transistor P2 is diode-connected in the first current path CP1. Therefore, when the potential of the test pad PAD is increased and V(bSN) is increased as shown in
The test current flowing through the second current path CP2 is determined based on the characteristic of the transistor P4 and the characteristic of the memory cell MC0 which is a target under the test. The transistor P4 is connected between the power source VDD and the sense node SN via the transistor P3. Therefore, the current flowing through the transistor P4 decreases when the potential of the sense node SN increases. On the contrary, the current flowing though the transistor P4 increases when the potential of the sense node SN decreases.
On the other hand, the target memory cell MC0 is connected between the sense node SN and the source potential VSL (ground potential). Therefore, the current flowing through the memory cell MC0 increases when the potential of the sense node SN increases. On the contrary, when the potential of the sense node SN decreases, the current flowing through the memory cell MC0 decreases. When the gate potential (word line voltage) of the memory cell MC0 is increased as shown by a broken-line arrowhead VWL in
The intersection between the characteristic of the transistor P4 and the characteristic of the memory cell MC0 expresses a test current flowing through the second current path CP2 and a voltage applied to the sense node SN. When the word line voltage VWL is changed from a low state VWL0 to a high state VWL2, the potential V(SN) of the sense node SN changes from Vsn_b to Vsn_a. A word line voltage VWL1 at which the potential V(SN) becomes equal to the reference voltage Vref can be detected.
Before time t1, the data “1” or “0” is written to the target memory cell MC concerned.
At time t1, the signal TS1 is activated, and the selection transistor P6 becomes in the on state. The selection transistor P5 maintains the off state. That is, in this case, the first test is executed to the memory cell of each column. In the second test, the states of the selection transistors P5 and P6 are reversed. Because the second test can be easily estimated, explanations thereof will be omitted.
At time t2, a signal TAIT0 is activated, and the transfer gate N5 becomes in the on state. Accordingly, the sense node SN is connected to the bit line BL0. The transfer gate N6 maintains the off state. In this case, the signal TFAIT is activated, and the transfer gates P7 and P8 become in the on state. Further, a voltage is applied to the selection word line WL0. At the beginning, the word line voltage VWL is set to the voltage VWL0 lower than the threshold voltage of the target memory cell MC0.
At time t3, a signal bLOADON is activated, and the current mirror circuit CM is driven. Accordingly, the reference current Iref flows from the power source VDD to the test pad PAD via the first current path CP1. At the same time, a test current according to the reference current Iref and the word line voltage VWL flows from the power source VDD to the memory cell MC0 via the second current path CP2. In the present embodiment, the first and the second current paths are replaced with each other. That is, the transistor P2 can be present in the first current path in some case and can be present in the second current path in other case. This is similarly applied to the transistor P4. Therefore, the mirror ratio of the transistor P2 to the transistor P4 needs to be one. Accordingly, a threshold current defining the threshold voltage becomes the reference current Iref. When Vref is equal to the drain voltage of the transistor P4 (or P2), the mirror ratio becomes the ratio of the test current flowing through the second current path to the reference current flowing through the first current path, that is, the ratio of the test current to the reference current (test current/reference current).
Because the word line voltage VWL is the voltage VWL0 lower than the threshold voltage of the target memory cell MC0, the potential of the sense node SN becomes higher than the reference voltage Vref of the sense node bSN.
At time t4, when a potential difference between the potentials of the sense nodes SN and bSN is sufficiently developed, both signals SAP and bSAN are activated. In this case, to prevent a through-current from SAP and bSAN to VSL, the signal FAIT0 is inactivated. To disconnect the sense amplifier S/A from the test circuit TC, the signal TFAIT is also inactivated. Accordingly, capacitances of the sense nodes SN and bSN become small, and the potential difference between the potentials of the sense nodes SN and bSN can be easily amplified.
By activating the signals SAP and bSAN, the latch circuits LC1 and LC2 are driven. Accordingly, the potential difference between the potentials of the sense nodes SN and bSN is amplified as shown in
At time t5, the column selection line CSL is activated. Accordingly, the transfer gates N1 and N2 become in the on state, and the sense node pair SN, bSN is connected to the DQ lines DQ, bDQ. Accordingly, the data latched into the sense node pair SN, bSN can be output to the outside via the DQ buffer.
At time t6, the signals SAP, bSAN, and TS1 are inactivated. Accordingly, a series of test operation of the word line voltage VWL0 ends.
Next, the test operation at t1 to t6 is repeated while increasing in stages the word line voltage VWL from VWL0. Because the data of FBC volatiles, the same data is rewritten to the memory cell MC, each time when the test operation at t1 to t6 is completed. When the word line voltage VWL exceeds the threshold voltage of the memory cell MC0, the data latched into the sense node pair SN, bSN are logically inverted. Accordingly, the threshold voltage of the memory cell MC can be measured. For example, when the word line voltage VWL exceeds the threshold voltage, the data latched into the sense node pair SN, bSN are inverted as shown in
The threshold voltage of the memory cell MC is executed both when the memory cell MC is the “0” cell and the “1” cell. The second test is also executed in a similar manner to that of the first test. Accordingly, the threshold voltage of the memory cell MC connected to the bit line bBL is also measured.
In the present embodiment, a threshold voltage of the memory cell MC is measured via the current mirror circuit CM. Accordingly, the test pad PAD can be made common to the test circuits TCs of all columns. That is, the test can be executed simultaneously to all columns, without the need of individually executing the test to each column. As a result, memory cells can be tested at a high speed, thereby decreasing the test time.
Because the first current path CP1 is made common to all columns within the memory cell array MCA, the second current paths CP2 and CP12 need to be provided in the bit lines BLj and bBLj, respectively.
The test circuit includes the current mirror circuit CM including the p-type transistors P1 to P4, P11, and P12, the selection transistors P5, P6, P13, and P14, and the transfer gates P7, P8. The transistors P1, P3, and P11 are switching elements that are on/off controlled by signals bLOADON, bLOADON_A, and bLOADON_B, respectively. The transistors P1, P3, and P11 become in the on state when a current is passed from the current mirror circuit CM.
The transistor P2 as a first transistor is present between the power source VDD via the transistor P1 and the test pad PAD. The transistor P4 as a second transistor is present between the power source VDD via the transistor P3 and the sense node SN. Further, the transistor P12 is present between the power source VDD via the transistor P11 and the sense node SN. That is, the transistor P2 is provided in the first current path CP1. The transistors P4 and P12 are provided in the second current paths CP2 and CP12, respectively.
In the second embodiment, the current paths CP1 and CP2 of the first and the second current paths respectively are not replaced with each other. Therefore, current driving capacities of the transistors P2 and P4 can be asymmetrical. That is, the current mirror circuit CM can have a predetermined mirror ratio. For example, when Vref is equal to the drain voltage of the transistor P4, the mirror ratio is a proportion of the current flowing through the second current path CP2 to the reference current flowing through the first current path CP1.
Current driving capacities of the transistors P2 and P12 can be asymmetrical. For example, when a size (W (channel width)/L (channel length)) of each of the transistors P4 and P12 is set to 1/10 of the size of the transistor P2, a current of 1/10 of the reference current flows through the transistors P4 and P12 when the memory cell MC is in the on state. Accordingly, a noise component contained in the test current can be minimized to 1/10 of the noise component contained in the reference current. Further, because the reference current Iref becomes large, the rising speed of the potential of the sense node and the like becomes fast. This leads to the shortening of the test time. The size (W/L) of the transistor P4 is equal to that of the transistor P12. In the second embodiment, the first current path CP1 is common to the second current paths CP2 and CP12 of all columns. Therefore, the reference current Iref flowing through the first current path CP1 is constant regardless of the number of columns and the number of the second current paths CP2.
The selection transistor P5 as a fourth transistor is connected between the gate of the transistor P4 and the gate of the transistor P2. The selection transistor P6 is connected between the sense node SN and the gate of the transistor P2. The selection transistor P5 is turned on at the time of testing the memory cell MC connected to the bit line BL, thereby connecting the gate of the transistor P2 and the test pad PAD to the gate of the transistor P4. Accordingly, the selection transistor P5 transmits the reference voltage to the gate of the transistor P4. The selection transistor P6 as a third transistor is turned on at the time of testing the memory cell MC connected to the bit line bBL, thereby connecting the gate of the transistor P2 and the test pad PAD to the sense node SN. Accordingly, the selection transistor P6 transmits the reference potential to the sense node SN.
The selection transistor P13 is connected between the gate of the transistor P12 and the gate of the transistor P2. The selection transistor P14 is connected between the sense node bSN and the gate of the transistor P2. The selection transistor P13 is turned on at the time of testing the memory cell MC connected to the bit line BL, thereby connecting the gate of the transistor P2 and the test pad PAD to the gate of the transistor P12. Accordingly, the selection transistor P13 transmits the reference voltage to the gate of the transistor P12. The selection transistor P14 is turned on at the time of testing the memory cell MC connected to the bit line BL, thereby connecting the gate of the transistor P2 and the test pad PAD to the sense node bSN. Accordingly, the selection transistor P14 transmits the reference voltage to the sense node bSN.
The second transistor (P4), the third transistors (P6, P14, . . . ), and the fourth transistors (P5, P13, . . . ) are provided corresponding to each sense node (each bit line).
A first test for testing the memory cells MCs connected to the bit lines BLj and a second test for testing the memory cells MCs connected to the bit lines bBLj are explained next. It is assumed that the transistors P1, P3, P7, and P8 are all in the on state during the execution of the first and the second tests.
In the first test, the selection transistors P5 and P14 become in the on state, and the selection transistors P6 and P13 become in the off state. Accordingly, a test current flows through the second current path CP2. In this case, the transistor P2 makes the first current path CP1 conductive, and passes the current (reference current) from the power source VDD to the test pad PAD. The test voltage generated by the reference current is transmitted to the sense node bSN via the transistor P14.
In the second test, the selection transistors P5 and P14 become in the off state, and the selection transistors P6 and P13 become in the on state. Accordingly, a test current flows through the second current path CP12. In this case, the test voltage generated by the reference current is transmitted to the sense node SN via the transistor P6. In the first test and the second test according to the second embodiment, the transistor P2 always makes the first current path CP1 conductive. In the first test, CP2 becomes the second current path. In the second test, CP12 becomes the second current path.
In the second embodiment, the first current path CP1 through which the reference current Iref is passed is common to the sense amplifier S/As of each column. Accordingly, the first current path CP1 can supply the constant stable reference voltage Vref to the sense amplifier S/A of each column. That is, the first current path CP1 can supply the reference voltage Vref having no variation to the sense amplifier S/A of each column. As a result, in the second embodiment, a threshold voltage of the memory cell MC can be measured in higher precision.
According to the second embodiment, the transistor P2 is made common to the sense amplifier S/A of each column. Therefore, the reference current Iref is constant, regardless of the number of columns, the number of the second current paths CP2, and the number of the sense amplifiers S/As.
When the mirror ratio (test current/reference current) is set smaller than one, the reference current Iref can be set larger than the test current. By setting the reference current Iref large, the rising speed of the potential of the sense node and the like can be made fast. This leads to the shortening of the test time.
In the second embodiment, a change of the size (W/L) of the transistor P2 gives smaller influence to the total size of the device than the change in the first embodiment.
The second embodiment can also achieve effects similar to those in the first embodiment.
The DQ buffer DQB is connected between the test pad PAD and the sense amplifier S/A, and includes latch circuits DQL1, DQL2, equalizing transistors P13 to P15, and a reading circuit RC. The latch circuits DQL1, DQL2 are cross-couple type dynamic latch circuits like the latch circuits LC1, LC2, respectively. However, the latch circuits DQL1, DQL2 are driven by signals QDRV, bQDRV, thereby latching data of the DQ line and the bDQ line. The transistors P13 to P15 are provided to equalize the DQ line and the bDQ line in the same potential during a standby period. Therefore, the transistors P13 to P15 become in the off state during the normal operation time or test time. The reading circuit RC is structured to output data latched into the DQ line and the bDQ line, to a reading unit RD at timing when a signal QENB is activated.
In the third embodiment, at time t2, the signal CEQ becomes inactive, thereby turning off the transistors P13 to P15. Accordingly, the DQ line and the bDQ line are electrically separated, and thereby ending the equalizing.
In the third embodiment, the signals SAP, bSAN remain inactive, and the latch circuits LC1 and LC2 of the sense amplifier S/A are not driven. On the other hand, the signals QDRV and bQDRV are activated at time t4. Accordingly, the latch circuits DQL1 and DQL2 of the DQ buffer DQB are activated, thereby amplifying a potential difference between the potentials of the DQ line and the bDQ line. Accordingly, the DQ buffer DQB detects data within the memory cell MC at the test time in place of the sense amplifier S/A, and outputs the data to the outside.
Further, at time t5, the signal QENB is activated, and the data latched into the DQ line and the bDQ line are output from the reading unit RD.
Other operations in the third embodiment are similar to those in the first embodiment.
One DQ buffer DQB is provided for plural sense amplifiers S/As. For example, one DQ buffer DQB is provided for each 256 columns or 512 columns. Therefore, by providing a test circuit TC corresponding to the DQ buffer DQB, a circuit scale of the total device can be made small. Further, the third embodiment can also achieve effects similar to those in the first embodiment.
Current driving capacities of the transistors P2 and P4 can be asymmetrical. That is, the current mirror circuit CM has a predetermined mirror ratio. Current driving capacities of the transistors P2 and P19 can be also asymmetrical. By making the mirror ratio small, the reference current Iref can be set large. By setting the reference current Iref large, the rising speed of the potential of the sense node and the like can be made fast. This leads to the shortening of the test time. The size (W/L) of the transistor P4 is equal to that of the transistor P19.
A first test and a second test in the fourth embodiment are substantially the same as those explained in the second embodiment. However, in the fourth embodiment, a test current flows to the memory cell concerned via the DQ line or the bDQ line. In other word, the second current path CP2 passes the test current to the memory cell MC via the DQ line, the sense node SN, and the bit line BL. The second current path CP12 passes the test current to the memory cell MC via the bDQ line, the sense node bSN, and the bit line bBL.
In the first test, because the test current is passed to the second current path CP2, the signal bLOADON_A is activated at the same timing as that of the signal bLOADON. In the first test, the signal bLOADON_B maintains the inactive state. In the second test, because the test current is passed to the second current path CP12, the signal bLOADON_B is activated at the same timing as that of the signal bLOADON. In the second test, the signal bLOADON_A maintains the inactive state. Other signal operations in the fourth embodiment are similar to those in the timing diagrams shown in
The fourth embodiment can achieve effects of both of the second and third embodiments.
In the first to the fourth embodiments, the transfer gates P7 and P8 can be n-type FETs.
In the above embodiments, the memory cell MC includes the n-type FETs. However, the memory cell MC can include p-type FETs. In this case, voltages applied to the selection word line and the bit line need to be inverted. Read data are logically inverted from the read data in the above embodiments.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2007-074783 | Mar 2007 | JP | national |
