Global Redundancy Column Selection Implementation for Boundary Faults in Memory Devices
By using a global redundant column selection repair circuit to dynamically adjust the data transmission path, the performance limitation caused by boundary faults in memory devices is solved, achieving rapid repair and efficient data transmission.
Patent Information
- Application Number
- CN202111587609.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2021-04-26
- Filing Date
- 2021-12-23
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2041-12-23
AI Technical Summary
When faced with boundary faults, existing memory devices may experience increased ECC processing complexity and specification constraints due to the implementation of local redundancy column selection, which limits the use of data lines and thus restricts the performance of the memory device.
A global redundant column selection repair circuit is adopted. Through the combination of logic gates and error correction circuits, the data transmission path is dynamically adjusted to ensure rapid repair of memory elements when a fault occurs. Parallel evaluation is performed in the error correction circuit to avoid the transmission of redundant data.
It enables rapid repair of faulty column planes in memory devices, meets boundary fault specifications, improves data transfer efficiency and memory device performance, and reduces ECC assessment latency.
Smart Images

Figure CN115249510B_ABST
Abstract
Description
Technical Field
[0001] Embodiments of this disclosure generally relate to the field of memory devices. More specifically, embodiments of this disclosure relate to the implementation of global redundancy column selection for boundary faults in memory devices. Background Technology
[0002] Memory devices are becoming increasingly complex due to the numerous possible failures in semiconductor manufacturing. To mitigate these potential failures, memory devices can be implemented using redundant memory elements that can be utilized when other memory elements fail. The defective memory element is deactivated, and the redundant memory element is activated. Memory access is then remapped from the defective memory element to other memory elements. In some cases, remapping may map the defective memory element to the redundant memory element. However, in other cases, multiple memory elements can be remapped by remapping the defective memory element to an intermediate memory element, (or another) intermediate memory element to the redundant memory element. Remapping can be performed within the CP using a column plane (CP) of memory elements having multiple (e.g., 2, 3, 4, or more) redundant columns, employing a Local Redundant Column Selection (LRCS) implementation to perform repairs within the CP. Alternatively, a Global Redundant Column Selection (GRCS) implementation in the memory device may include a global column plane used instead of a faulty column selection. In some embodiments, GRCS implementations may have better yields and smaller die sizes than some LRCS implementations.
[0003] However, GRCS implementations can utilize data multiplexing processed by Delayed Error Correction Code (ECC). Furthermore, GRCS implementations can be constrained by ECC specifications. For example, a memory device boundary fault specification can specify how many bits on the die (e.g., 1 bit) and how many bits at the system level (e.g., multiple bits) are correctable. Thus, the specification can indicate a first number (e.g., 1) of data lines (DQs) for all burst bits and a second number (e.g., 2) of DQs for half-burst bits that are correctable. Additionally, the specification can indicate where bit errors are allowed. For example, the specification can indicate that all faulty data is within a single DQ or only within the two DQs of a half-burst. Furthermore, the memory device specification can limit the number of DQs on the subword line (SWL) (e.g., 2). This limitation can prevent the safe use of certain GRCS implementations that might result in more DQs than the number on the SWL.
[0004] The embodiments disclosed herein may address one or more of the problems set forth above. Summary of the Invention
[0005] One aspect of this disclosure relates to an electronic device comprising a plurality of memory banks; and a repair circuit configured to remap data from the plurality of memory banks to repair memory elements of the plurality of memory banks in the event of a fault, wherein the repair circuit comprises: a logic gate configured to: receive an output from a memory bank of the plurality of memory banks; receive a fault signal indicating whether a corresponding memory element of the plurality of memory banks has failed; and transmit an output, wherein the value of the output is at least partially based on the fault signal; and an error correction circuit configured to receive the output via the logic gate; and a multiplexer configured to: receive the output from the memory banks; receive a repair value; and selectively output the output or the repair value from the repair circuit as the output of the repair circuit.
[0006] Another aspect of this disclosure relates to a repair circuit for a memory device, comprising: a plurality of logic gates, each configured to: receive a corresponding output from a corresponding amplifier among a plurality of amplifiers; receive a corresponding fault signal indicating whether to repair a corresponding memory element; and selectively transmit a corresponding output or a default value to an electronic correction circuit, at least in part based on the corresponding fault signal; and a plurality of multiplexers, each configured to: receive a corresponding output from a corresponding amplifier; receive a repair value; and selectively output a repair value or a corresponding output on a corresponding data line among a plurality of data lines.
[0007] Another aspect of this disclosure relates to a method comprising: receiving, at a logic gate, an output of a data sense amplifier from a data sense amplifier; receiving, at a logic gate, an indication of whether a column plane of a memory bank is repaired; selectively transmitting an output or a default value from the logic gate to an error correction circuit, at least in part based on the indication; determining, in the error correction circuit, whether a bit error exists in the output; receiving, at a multiplexer, the output of the data sense amplifier from a data sense amplifier; receiving a repair value at the multiplexer; receiving an indication at the multiplexer; and selectively outputting, based on the indication, the output of the data sense amplifier or the repair value, wherein at least a portion of the determination is performed before the selective output is completed. Attached Figure Description
[0008] Figure 1 A simplified block diagram illustrating certain features of a memory device with repair circuitry according to embodiments of the present disclosure;
[0009] Figure 2 According to embodiments of this disclosure Figure 1 The mapping of data lines (DQ) in the memory bank of the memory device;
[0010] Figure 3 For embodiments of the present disclosure having a global column plane (GCP) Figure 2 A representation of part of DQ;
[0011] Figure 4 For the use of GCP according to embodiments of this disclosure Figure 1 A shift-based GRCS implementation of the repair circuit;
[0012] Figure 5 For the shift repair using GCP according to embodiments of this disclosure Figure 3 A representation of part of DQ;
[0013] Figure 6 For the use of alternative GRCS implementations according to embodiments of this disclosure Figure 1 The circuit diagram of the repair circuit;
[0014] Figure 7 To achieve the effect of not making repairs according to embodiments of this disclosure. Figure 6 The repair circuit output Figure 2 A portion of the DQ representation; and
[0015] Figure 8 In order to perform repairs according to embodiments of this disclosure Figure 6 The repair circuit output Figure 2 A representation of part of DQ;
[0016] Figure 9 For use according to embodiments of this disclosure Figure 4 A flowchart of data reading implemented using shift-based GRCS; and
[0017] Figure 10 For use according to embodiments of this disclosure Figure 6 The flowchart for data reading in the alternative implementation of GRCS. Detailed Implementation
[0018] One or more specific embodiments will be described below. To provide a concise description of these embodiments, not all features of the actual implementation are described in the specification. It should be understood that, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developer's specific goals, such as complying with system-related and business-related constraints, which may vary from one implementation to another. Furthermore, it should be understood that such development work can be complex and time-consuming; however, these are routine tasks in design, construction, and manufacturing for those skilled in the art to which this disclosure pertains.
[0019] A repair circuit can be used to repair faulty column planes using a global column plane. The repair circuit receives the output from a data sense amplifier (DSA) and transmits the output (e.g., simultaneously and / or in parallel) to both a multiplexer and logic gates. The logic gates can be used to connect the output to error correction circuitry for performing error correction code (ECC) evaluation and / or correction. Data from the repaired column plane may not be used for actual data. Therefore, it is not necessary to evaluate the contents of these column planes for bit errors in the error correction circuitry. Instead, the logic gates can be used to set the corresponding contents to a default value (e.g., zero) based on a signal indicating whether the column plane has been repaired. The same signal can be used to control whether the DSA output is output from the corresponding multiplexer as a data input-output (DQ). If the column plane is not repaired, then the output is transmitted. If the column plane is repaired, then data from the global column plane (GCP) can be output from the multiplexer instead of from the DSA. Furthermore, the GCP can be evaluated separately when transmitted from the corresponding DSA corresponding to the GCP. If data from the GCP is used (e.g., to repair any column plane), then the GCP logic gate can transmit the repaired value of the DSA to the error correction circuitry. If data from the GCP is not used, then the GCP logic gate can set the GCP to its default value.
[0020] Now turn to the diagrams. Figure 1 A simplified block diagram illustrating certain characteristics of the memory device 10. Specifically, Figure 1 The block diagram is a functional block diagram illustrating certain functionalities of the memory device 10. According to one embodiment, the memory device 10 may be a fifth-generation dual data rate synchronous dynamic random access memory (DDR5 SDRAM) device. Compared to previous generations of DDR SDRAM, various features of DDR5 SDRAM allow for reduced power consumption, more bandwidth, and greater storage capacity.
[0021] Memory device 10 may include several memory groups 12. Memory groups 12 may be, for example, DDR5 SDRAM memory groups. Memory groups 12 may be disposed on one or more chips (e.g., SDRAM chips) arranged on a dual in-line memory module (DIMM). As will be understood, each DIMM may contain several SDRAM memory chips (e.g., x4, x8, or x16 memory chips). Each SDRAM memory chip may contain one or more memory groups 12. Memory device 10 represents a portion of a single memory chip (e.g., an SDRAM chip) having several memory groups 12. For DDR5, memory groups 12 may be further arranged to form clusters. For example, for 8 gigabit (Gb) DDR5 SDRAM, the memory chip may contain 16 memory groups 12 arranged in 8 clusters, each cluster containing 2 memory groups. For example, for 16Gb DDR5 SDRAM, the memory chip may contain 32 memory groups 12 arranged in 8 clusters, each cluster containing 4 memory groups. Depending on the application and design of the entire system, various other configurations, organization and sizes of the memory group 12 on the memory device 10 may be utilized.
[0022] The memory device 10 may include a command interface 14 and an input / output (I / O) interface 16. The command interface 14 is configured to provide several signals (e.g., signal 15) from an external device (not shown), such as a processor or controller. The processor or controller may provide various signals 15 to the memory device 10 to facilitate the transfer and reception of data to be written to or read from the memory device 10.
[0023] As will be understood, command interface 14 may include several circuits, such as clock input circuitry 18 and command address input circuitry 20, to ensure proper handling of signal 15, for example. Command interface 14 may receive one or more clock signals from an external device. Generally, Double Data Rate (DDR) memory utilizes a differential pair of system clock signals, referred herein as the true clock signal (Clk_t) and the complementary clock signal (Clk_c). The positive clock edge of DDR refers to the point where the rising true clock signal Clk_t crosses the falling complementary clock signal Clk_c, while the negative clock edge indicates the transition of the falling true clock signal Clk_t and the rising of the complementary clock signal Clk_c. Commands (e.g., read commands, write commands, etc.) are typically input on the positive edge of the clock signal, and data is transmitted or received on both the positive and negative clock edges.
[0024] Clock input circuit 18 receives a true clock signal (Clk_t) and a complementary clock signal (Clk_c) and generates an internal clock signal CLK. The internal clock signal CLK is supplied to an internal clock generator 30, such as a delay-locked loop (DLL) circuit. The internal clock generator 30 generates a phase-controlled internal clock signal LCLK based on the received internal clock signal CLK. The phase-controlled internal clock signal LCLK is supplied to, for example, I / O interface 16 and used as a timing signal to determine the output timing for reading data.
[0025] The internal clock signal CLK can also be provided to various other components within the memory device 10 and can be used to generate various additional internal clock signals. For example, the internal clock signal CLK can be provided to the command decoder 32. The command decoder 32 can receive command signals from the command bus 34 and can decode the command signals to provide various internal commands. For example, the command decoder 32 can provide command signals to the internal clock generator 30 via bus 36 to coordinate the generation of the phase-controlled internal clock signal LCLK. The phase-controlled internal clock signal LCLK can be used, for example, to time data via the I / O interface 16.
[0026] Furthermore, the command decoder 32 can decode commands such as read commands, write commands, mode register setting commands, and activation commands, and provides access to specific memory groups 12 corresponding to the commands via bus path 40. As will be understood, the memory device 10 may include various other decoders, such as row decoders and column decoders, to facilitate access to the memory groups 12. In one embodiment, each memory group 12 includes a group control block 22 that provides the necessary decoding (e.g., row decoders and column decoders) and other features, such as timing control and data control, to facilitate the execution of commands to and from the memory groups 12. For example, the group control block 22 may include a repair circuit 23 that can remap data in the memory groups 12 using redundant memory elements. For example, the repair circuit 23 can be used to implement Local Redundancy Column Select (LRCS) repair and / or Global Redundancy Column Select (GRCS) repair.
[0027] Memory device 10 performs operations such as read and write commands based on command / address signals received from an external device, such as a processor. In one embodiment, the command / address bus may be a 14-bit bus for receiving command / address signals (CA<13:0>). The command / address signals are timed to command interface 14 using clock signals (Clk_t and Clk_c). The command interface may include command address input circuitry 20 configured to receive and transmit commands via, for example, a command decoder 32 to provide access to memory bank 12. Additionally, command interface 14 may receive a chip select signal (CS_n). The CS_n signal enables memory device 10 to process commands input on the CA<13:0> bus. Access to a specific bank 12 within memory device 10 is encoded as a command on the CA<13:0> bus.
[0028] Additionally, command interface 14 can be configured to receive several other command signals. For example, a command / address (CA_ODT) signal on the die termination can be provided to facilitate proper impedance matching within memory device 10. For example, a reset command (RESET_n) can be used during power-up to reset command interface 14, status registers, state machines, etc. Command interface 14 can also receive a command / address inverted (CAI) signal, which can be provided to invert the state of command / address signals CA<13:0> on the command / address bus. A mirror (MIR) signal can also be provided to facilitate mirroring functionality. Depending on the configuration of multiple memory devices in a particular application, the MIR signal can be used to multiplex the signal so that it can be interchanged to implement a particular route to the memory device 10. Various signals can also be provided to facilitate testing of memory device 10, such as a test enable (TEN) signal. For example, the TEN signal can be used to put memory device 10 into test mode for connectivity testing.
[0029] Command interface 14 can also be used to provide a warning signal (ALERT_n) to the system processor or controller in response to certain detectable errors. For example, the warning signal (ALERT_n) can be transmitted from memory device 10 upon detection of a cyclic redundancy check (CRC) error. Other warning signals may also be generated. Furthermore, the bus and pins used for transmitting the warning signal (ALERT_n) from memory device 10 can be used as input pins during certain operations, such as the connectivity test mode performed using the TEN signal as described above.
[0030] Data can be sent to and from memory device 10 by transmitting and receiving data signals 44 via I / O interface 16 using the command and timing signals discussed above. More specifically, data can be sent to or retrieved from memory group 12 via data path 46, which includes multiple bidirectional data buses. Data I / O signals (commonly referred to as DQ signals) on the DQ line are typically transmitted and received on one or more bidirectional data buses. For some memory devices, such as DDR5 SDRAM memory devices, I / O signals can be divided into upper and lower bytes. For example, for x16 memory devices, I / O signals can be divided into upper and lower I / O signals corresponding to upper and lower bytes of, for example, data signals (e.g., DQ<15:8> and DQ<7:0>).
[0031] To allow higher data rates within memory device 10, some memory devices, such as DDR memory devices, may utilize data strobe signals, commonly referred to as DQS signals. The DQS signals are driven by an external processor or controller transmitting data (e.g., for write commands) or by memory device 10 itself (e.g., for read commands). For read commands, the DQS signals are essentially additional DQ signals with a predetermined pattern. For write commands, the DQS signals are used as clock signals to capture the corresponding input data. Similar to clock signals (Clk_t and Clk_c), differential pairs (DQS_t and DQS_c) of data strobe (DQS) signals can be provided as data strobe signals to provide differential pair signaling during read and write operations. For certain memory devices, such as DDR5 SDRAM memory devices, the differential pair of the DQS signal can be divided into upper and lower data strobe signals (e.g., UDQS_t and UDQS_c; LDQS_t and LDQS_c), which correspond to the upper and lower bytes of data, for example, sent to and from memory device 10.
[0032] The impedance (ZQ) calibration signal can also be provided to the memory device 10 via I / O interface 16. The ZQ calibration signal is provided to a reference pin and used to tune the output driver and ODT values by adjusting the pull-up and pull-down resistors of the memory device 10 by changing the process, voltage, and temperature (PVT) values. Because PVT characteristics can affect the ZQ resistor value, the ZQ calibration signal is provided to the ZQ reference pin for adjusting the resistance to calibrate the input impedance to a known value. As will be understood, a precision resistor is typically coupled to the ZQ pin on the memory device 10 to the external GND / V of the memory device 10. SS Between. This resistor acts as a reference for adjusting the drive strength of the internal ODT and IO pins.
[0033] Additionally, a loopback signal (LOOPBACK) can be provided to memory device 10 via I / O interface 16. The loopback signal can be used during test or debugging phases to set memory device 10 into a mode where signals loop back through memory device 10 via the same pin. For example, the loopback signal can be used to set memory device 10 to test its data output (DQ). The loopback may include both data and strobe pins or may only include the data pin. This is typically intended for monitoring data captured by memory device 10 at I / O interface 16.
[0034] As will be understood, various other components, such as power supply circuitry (for receiving external Vdd and Vss signals), mode registers (for defining various modes of programmable operation and configuration), read / write amplifiers (for amplifying signals during read / write operations), and temperature sensors (for sensing the temperature of the memory device 10), may also be incorporated into the memory device 10. Accordingly, it should be understood that only [the following is provided] Figure 1 The block diagram is used to highlight certain functional features of the memory device 10 to aid in the subsequent detailed description.
[0035] With the aforementioned ideas in mind, Figure 2 Mapping 50 is shown for the data lines (DQs) in an embodiment of DDR5 x8 DQ mapping for memory bank 12. Mapping 50 is for illustrative purposes, but other embodiments may include other mappings and / or a different number of DQs per group. As illustrated, memory bank 12 includes column planes 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, 72, 74, 76, 78, 80, 82, and 84, collectively referred to as column planes 52-84. Some embodiments may have a different number of column planes per group and / or per memory bank. Each of column planes 52-84 is associated with a DQ for the corresponding column address and even / odd sense amplifier. For example, column planes 52 and 54 correspond to DQ7. Column planes 56 and 58 correspond to DQ5, column planes 60 and 62 correspond to DQ6, and column planes 64 and 66 correspond to DQ4. Column plane 68 corresponds to an error control code (ECC). Column planes 70 and 72 correspond to DQ3, column planes 74 and 76 correspond to DQ1, column planes 78 and 80 correspond to DQ2, and column planes 82 and 84 correspond to DQ0.
[0036] Figure 3 For corresponding Figure 2The corresponding subword lines (SWLs) 92, 94, and 96 of the DQ are graphically represented 90. As illustrated, the DQs can be strategically positioned so that if an SWL fails, the boundary fault specification is still met. Graphical representation 90 shows the association of SWLs 92, 94, and 96 with column planes 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, and 72 without repairing column planes 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, and 72. As illustrated, SWL 92 corresponds to DQ7 and DQ5. For example, SWL 92 can be coupled to... Figure 2 The mapping 50 corresponds to column planes 52, 54, 56, and 58. Similarly, SWL 94 corresponds to DQ6 and DQ4. For example, SWL 94 can be coupled to... Figure 2 The mapping 50 has column planes 60, 62, 64, and 66. Additionally, SWL 96 corresponds to a parity bit (e.g., ECC), a global column plane (GCP), and DQ3. In embodiments where the specification of memory device 10 does not specify isolating the ECC parity bit from the GCP and / or DQ (e.g., DQ3), SWL 96 may correspond to both the parity bit and the GCP. In some embodiments, at least some DQs (e.g., DQ7, DQ0, etc.) may have a half-length SWL of some WLs (e.g., even WLs).
[0037] In order to perform repair on memory device 10, repair circuit 23 can be used to shift / remap data to perform repair. Figure 4This is a schematic diagram of a shift-based implementation of repair circuit 100, also known as repair circuit 23. Repair circuit 100 repairs a memory group 102 (e.g., memory group 12) comprising redundant memory cell groups (MCGs) 104 and MCGs 106, 108, 110, and 112. Redundant MCGs 104 and 106, 108, 110, and 112 may each have one or more corresponding column planes. Redundant MCGs 104 and 106, 108, 110, and 112 can output corresponding data 114, 116, 118, 120, and 122 to a data sense amplifier 124. The data sense amplifier 124 then transmits the data 116, 118, 120, and 122 to corresponding multiplexers 126, 128, 130, and 132. To accommodate column plane repair, each multiplexer 126, 128, 130, and 132 can receive data 114 from redundant MCG 104. Alternatively, as illustrated, data from multiple column planes can be shifted to reduce shift and / or routing distances. In other words, each multiplexer 126, 128, 130, and 132 can be used to select between two different MCGs / column planes. For example, multiplexer 126 can receive data 114 and data 116, multiplexer 128 can receive data 116 and data 118, multiplexer 130 can receive data 118 and 120, and multiplexer 132 can receive data 122 and data 134 received from an MCG (not shown) adjacent to MCG 112.
[0038] Selected inputs from each of multiplexers 126, 128, 130, and 132 can be used as corresponding DQ outputs. For example, the output of multiplexer 126 can correspond to DQ0 136, the output of multiplexer 128 can correspond to DQ1 138, the output of multiplexer 130 can correspond to DQ2 140, and the output of multiplexer 132 can correspond to DQ7 140.
[0039] Control logic circuitry 144 can be used to generate and / or output one or more control signals 146 to control the selection of inputs to multiplexers 126, 128, 130, and 132. For example, shifting can be used to shift data in the repaired column plane 147 of MCG 110. Control logic circuitry 144 can receive an indication of and / or determine the position of the data shift boundary 148 corresponding to the repaired column plane 147. All read data to the left of the data shift boundary 148 in shift region 150 is shifted one column plane length to the left for multiplexing, while data to the right of the data shift boundary 148 is in the unshifted non-shift region 152. Therefore, data in DQ2 140 is shifted from MCG 108 as data 118 instead of data 120 from MCG 110 or data 114 from redundant MCG 104, while data in DQ7 142 is transmitted from MCG 112 as data 122. Due to shift repair, data 118 from MCG 108 is transmitted as DQ2 140, data 116 from MCG 106 is transmitted as DQ1 138, and data 114 from redundant MCG 104 is transmitted as DQ0 136. Each of the DQs can then be transmitted to other circuits, such as ECC circuitry and / or the host device. Therefore, ECC evaluation is delayed until multiplexing is performed before the DQs can be evaluated for bit errors. In other words, the repair circuitry 100 can delay ECC calculation.
[0040] Although the shifting scheme of the repair circuit 100 may be able to repair the repaired column plane 147 through a shortened route, the output of the repair circuit 100 may conflict with the specifications of the memory device 10 (e.g., ECC requirements). For example, Figure 5 To correspond to the repair after using the repair circuit 100 Figure 2 The corresponding DQ sub-word lines (SWL) 92, 94, and 96 are graphically represented as 160. In other words, Figure 5 The graphic representation 160 can be displayed on the Figure 3 The corresponding representations are SWLs 92, 94, and 96 after the repair of column plane 58. As explained, DQ5 corresponding to column plane 58 is shifted from position 162 using repair circuit 100. Alternatively, the portion of DQ5 is shifted from SWL 92 to SWL 94. Similarly, the data corresponding to DQ6, DQ4, and the parity bit are also shifted until the parity bit is shifted to the position assigned to the parity bit. Figure 2 The graphic representation is within the GCP space in 90. However, due to the shift, portions of DQ6, DQ5, and DQ4 are on SWL 94. This situation may occur outside the boundary fault specifications of memory device 10 (e.g., limitations on permissible values due to ECC evaluation).
[0041] In order to maintain compliance with the specifications of memory device 10, Figure 6 The repair circuit 180 can be used as Figure 1 An alternative implementation of the repair circuit 23. As illustrated, the repair circuit 180 receives data from data sense amplifiers (DSA) 182, 184, 186, and 188 that sense data stored in memory block 12. Although only four DSAs are shown in the repair circuit 180, embodiments of the repair circuit 180 may have fewer or more DSAs. For example, a DSA may exist for each bit in a column plane, with each illustrated DSA corresponding to a respective column plane. Furthermore, each DSA to the left of the GCP and parity DSA may correspond to the column plane to the left of column plane 68, while each DSA to the right of the GCP and parity DSA may correspond to the column plane to the right of column plane 68. Additionally, a number of data bit DSAs (e.g., 128) and a second number of GCP bit DSAs (e.g., 8 bits) may exist. Thus, each DSA may represent multiple DSAs and / or multiple bits may be output to multiple logic gates. Furthermore, the number of parity bits may vary depending on the ECC scheme.
[0042] In addition, each DSA has a corresponding output. For example, DSA 182 has an output 190 that outputs to AND gate 192. AND gate 192 also receives a column plane fault signal 194 indicating whether the corresponding column plane has been repaired. For example, if the column plane has been repaired, then column plane fault signal 194 may be low, thereby causing AND gate 192 to zero out output 190 before transmitting data to ECC circuit 195 for evaluating bit errors. Output 190 is also transmitted from DSA 182 to multiplexer 196 to determine whether output 190 or Global Column Plane (GCP) output 197 is transmitted as DQ output 198 for DSA 182. DSA 182 is configured to transmit output 190 to multiplexer 196 and AND gate 192 simultaneously and / or in parallel. (See below for more details.) Figure 10 As discussed, by sending output 190 (or zero) instead of waiting for the multiplexer to complete, the ECC calculation on output 190 can be performed faster than multiplexing before ECC evaluation. If a repair has occurred, a faster replacement of output 190 with zeros can be performed compared to using a multiplexer between DSA182 and ECC circuit 195, because a simple repair signal can be combined with output 190 via AND gate 192. This zeroing is possible because the repair uses data from GCP output 197 instead of data from output 190. As illustrated, GCP output 197 can be independently transmitted from DSA 186 to ECC circuit 195. Therefore, for ECC evaluation, the zeroed output 190 does not need to be replaced by GCP output 197. Alternatively, output 190 can be set to a default value (e.g., all zeros).
[0043] The DSA 184 outputs a parity output 200, which, along with a parity column plane fault signal 202, is transmitted to an AND gate 204. If a column plane used for ECC fails and / or is optionally disabled, the parity column plane fault signal 202 (similar to column plane fault signal 194) can be used to zero out the parity output 200 before it is transmitted to the ECC circuit 195 for ECC evaluation.
[0044] The DSA 186 output can be used to replace data from other outputs of other DSAs to repair the column plane GCP output 197. The column plane fault signal 206 indicates whether the GCP output 197 is used to repair any column plane faults. In other words, if no repair is performed, then the GCP output 197 may not be used. In such cases, the GCP output 197 can be zeroed via AND gate 208 using the column plane fault signal 206 before being transmitted to the ECC circuit 195. In other words, when repair is not used on the memory bank, the repair logic gate (e.g., an AND gate) can be used to zero the repair value before it is transmitted to the ECC circuit 195. However, when repair is used, the GCP output 197 can be transmitted to the ECC circuit 195 from the repair logic gate (e.g., AND gate 208) rather than from the location where the contained data is used, because this replacement can be performed at least partially in parallel with the ECC evaluation in the multiplexer.
[0045] Except for the different column planes associated with the right side of DSA 188 and DSA 184 and 186, DSA 188 functions similarly to DSA 182. Specifically, DSA 188 can output output 210. A corresponding column plane fault signal 212 indicates whether the column plane corresponding to DSA 188 and output 210 has been repaired to use GCP output 197. If repair has been performed, the data being used is then transmitted to ECC circuit 195 as GCP output 197. In such cases, it is not necessary to check output 210 for ECC evaluation. Therefore, output 210 can be zeroed using an AND gate and column plane fault signal 212. Output 210 is also transmitted to multiplexer 218, which uses column plane fault signal 212 to control whether output 210 is output from multiplexer 218 as corresponding DQ 220. In some embodiments, repair may be limited to one column plane at a time.
[0046] Although the foregoing discussion of zeroing out unused data before it is transmitted to ECC circuit 195 is discussed, some embodiments may utilize alternative default values (e.g., 1) when the corresponding column plane has been repaired and the unrepaired output will not be used in ECC circuit 195.
[0047] Figure 7 For corresponding Figure 2The corresponding subword lines (SWLs) 222, 224, and 226 of the DQ are graphically represented 221. Graphical representation 221 shows the association of SWLs 222, 224, and 226 with column planes 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, and 72 without repair of column planes 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, and 72. As illustrated, SWL 222 corresponds to DQ7 and DQ5. For example, SWL 222 can be coupled to... Figure 2 The mapping 50 corresponds to column planes 52, 54, 56, and 58. Similarly, SWL 224 corresponds to DQ6 and DQ4. For example, SWL 224 can be coupled to... Figure 2 The mapping 50 corresponds to column planes 60, 62, 64, and 66. Additionally, SWL 226 corresponds to parity bits (e.g., ECC), the global column plane (GCP), and DQ3. In embodiments where the specification of memory device 10 does not specify isolation of the ECC parity bits from GCP and / or DQ (e.g., DQ3), SWL 226 may correspond to both parity bits and GCP. Since GCP output 197 is not used, it has been used... Figure 5 The repair circuit 180 resets it to zero, as indicated by zero 228. Although a portion of the DQ is shown in... Figure 7 However, additional DQ and / or column planes in a similar arrangement may be included in the memory device 10.
[0048] Figure 8 For transmission to ECC circuit 195, corresponding to Figure 2 Graphical representation 230 shows the corresponding sub-word lines (SWL) 232, 234, and 236 of DQ. Graphical representation 221 shows the association of SWL 232, 234, and 236 with column planes 52, 54, 56, 58, 60, 62, 64, 66, 68, 70, and 72 when column plane 58 is repaired. As illustrated, SWL 232 corresponds to DQ7 and DQ5. However, the previous repair with... Figure 7 The column plane 58 associated with DQ5 is described in the diagram. Therefore, as described, the output corresponding to column plane 58 (e.g., output 190) is zeroed using a corresponding AND gate (e.g., AND gate 192) and a corresponding column plane fault signal (e.g., column plane fault signal 194).
[0049] SWL 234 corresponds to DQ6 and DQ4. For example, SWL 234 can be coupled to... Figure 2The mapping 50 corresponds to column planes 60, 62, 64, and 66. Additionally, SWL 236 corresponds to parity bits (e.g., ECC), GCP, and DQ3. However, since column plane 58 has been repaired, redundant memory in the GCP is used to store data already stored in column plane 58. Therefore, the GCP at position 240 is transferred to ECC circuit 195 containing data used as DQ5. Specifically, the GCP output (e.g., GCP output 197) is used to replace the repaired output (e.g., output 190) as the corresponding portion of DQ5 using a corresponding multiplexer (e.g., multiplexer 196). This portion of DQ5 is evaluated in parallel with multiplexing in ECC circuit 195, allowing ECC circuit 195 to perform evaluation without waiting for multiplexing to complete. Furthermore, although portions of the data from DQ5 are evaluated in ECC circuit 195, due to the presence of parity bits and GCP and DQ3, SWL will not carry too many (e.g., 3) different DQs. For example, SWL232 carries only DQ7 and DQ5, SWL 234 carries only DQ6 and DQ4, and SWL 236 carries only DQ5 and DQ3. This repair result enables memory device 10 to operate within the boundary fault specification.
[0050] Furthermore, as mentioned above, by removing multiplexing from the critical path of the ECC circuit 195, evaluation data using the ECC circuit 195 can be performed more quickly. For example, Figure 9 For demonstration purposes Figure 4 A flowchart of the timing process in the read process 250 of the repair circuit 100 is provided. As illustrated, data is acquired from the DSA via the main IO (MIO) line during MIO cycle 252. A pulse based on the data sense amplifier clock (ClkDSA) 254 is used to determine that the output from the DSA has been latched. During multiplexer cycle 256, multiplexers (e.g., multiplexers 126, 128, 130, and 132) select which outputs are transmitted for ECC evaluation. After the multiplexers transmit the selected outputs, the ECC circuit evaluates the parity error of the transmitted outputs in ECC cycle 258. A pulse (e.g., ErrDec 260) may indicate that the ECC calculation has been completed and / or indicate whether a bit error has been detected / corrected. During load cycle 262, verified (and / or corrected) data is loaded onto the drive bus, which drives the data output after confirming the drive enable signal 264 (e.g., the DrEn signal).
[0051] Figure 10 For demonstration purposes Figure 6A flowchart illustrating the timing of the read process 280 of the repair circuit 180 is provided. As illustrated, data is acquired from the DSA via the main IO (MIO) line during MIO cycle 282. Pulses based on the data sense amplifier clock (ClkDSA) 284 are used to determine if the output from the DSA has been latched. During multiplexer cycle 286, the multiplexer (e.g., multiplexers 196, 218) selects which outputs to transmit for use as the corresponding DQ. As the multiplexer transmits the selected outputs, the ECC circuit 195 evaluates the outputs also transmitted to the multiplexer, allowing the ECC circuit 195 to evaluate the outputs in ECC cycle 288 without waiting for the multiplexer to complete the selection and transmission of the outputs. In other words, ECC cycle 288 may at least partially overlap with multiplexer cycle 286, allowing ECC cycle 288 to be performed after the start of the read process 280. Figure 9 The ECC cycle 258 begins earlier. A pulse (e.g., ErrDec 290) can indicate that the ECC calculation has been completed and / or indicate whether a parity bit error has been detected. Verified (and / or corrected) data is loaded onto the drive bus during load cycle 292, which drives data output after confirming drive enable signal 294. Because ECC cycle 288 can be closer to the start of read process 280 relative to the start of read process 250 than ECC cycle 258, read process 280 can complete faster than read process 250. Furthermore, read process 280 avoids associating too much DQ with SWL that may exist in shift-based repair implementations.
[0052] While this disclosure allows for various modifications and alternatives, specific embodiments have been shown by way of example in the drawings and described in detail herein. However, it should be understood that this disclosure is not intended to limit it to the specific forms disclosed. In fact, this disclosure is intended to cover all modifications, equivalents, and alternatives that fall within the spirit and scope of this disclosure as defined by the appended claims. For example, an AND gate may be replaced by a NAND gate with signals having opposite polarities. Alternatively or additionally, an inverter (e.g., an inverting amplifier) may be used while maintaining the spirit and scope of this disclosure as defined by the appended claims.
[0053] The techniques presented and claimed herein are referenced and applied to tangible objects and specific examples of practical nature, which demonstrably improve the technical field of the invention and are therefore not abstract, intangible, or purely theoretical. Furthermore, if any claim appended to this specification contains one or more elements expressed as "a component for [performing] [the function]..." or "a step for [performing] [the function]...", such elements are expected to be interpreted in accordance with 35 USC112(f). However, for any claim containing elements specified in any other way, such elements are not expected to be interpreted in accordance with 35 USC112(f).
Claims
1. An electronic device comprising: a plurality of memory banks; as well as A repair circuit configured to remap data from the plurality of memory groups to repair memory elements of the plurality of memory groups in the event of a failure, wherein the repair circuit includes: The logic gates are configured as follows: Receive output from the memory group among the plurality of memory groups; Receive a fault signal indicating whether a corresponding memory element in the plurality of memory groups has failed, wherein the logic gate is configured to return the output to zero when the fault signal indicates that the memory element has failed; and Transmit the output, wherein the value of the output is at least partially based on the fault signal; and An error correction circuit configured to receive the output via the logic gate; The multiplexer is configured as follows: Receive the output from the memory group; Receive repair value; and The repair circuit selectively outputs either the output or the repair value as its output, wherein zeroing the output occurs before transmission to the error correction circuit without altering the output transmitted to the multiplexer.
2. The electronic device of claim 1, wherein the error correction circuit is configured to detect whether data from the plurality of memory groups contains errors before or during the multiplexer selectively outputs the output or the repair value.
3. The electronic device of claim 1, wherein the repair circuit includes a repair logic gate configured to zero out the repair value before transmission to the error correction circuit without altering the repair value transmitted to the multiplexer.
4. The electronic device of claim 3, wherein the repair logic gate includes an AND gate configured to receive the repair value and a repair signal indicating whether repair will be performed for any of the plurality of memory groups, and wherein the AND gate is configured to reset the repair value to zero at least in part based on the repair signal.
5. The electronic device of claim 4, wherein the memory element comprises a column plane of the memory group in the plurality of memory groups.
6. The electronic device of claim 5, wherein the repair value is received from a global column plane, and wherein the repair signal indicates that any column plane of the memory group has failed, while the fault signal indicates a fault state of a specific column plane.
7. The electronic device of claim 1, wherein the logic gate includes an AND gate.
8. The electronic device of claim 1, further comprising a data sensing amplifier configured to receive an indication of the output from one of the plurality of memory groups and to transmit the output to the logic gate and the multiplexer.
9. The electronic device of claim 8, wherein the data sensing amplifier is configured to transmit the output simultaneously to the logic gate and the multiplexer.
10. The electronic device of claim 1, wherein the error correction circuit is configured to detect, at least in part, whether data from the plurality of memory groups contains errors while the multiplexer selects the output or repair value.
11. A repair circuit for a memory device, comprising: Multiple logic gates, each configured as follows: Receive the corresponding output from the corresponding amplifier among multiple amplifiers; Receive an indication of whether to repair the corresponding fault signal of the corresponding memory element; and The corresponding output or default value is selectively transmitted to the electronic correction circuit based at least in part on the corresponding fault signal, wherein the default value includes all zeros formed at least in part based on the corresponding fault signal; as well as Multiple multiplexers, each configured as follows: Receive the corresponding output from the corresponding amplifier; Receive repair value; as well as The repair value or the corresponding output is selectively output on the corresponding data line among multiple data lines.
12. The repair circuit of claim 11, wherein receiving the corresponding output from the corresponding amplifier among the plurality of amplifiers includes receiving the corresponding output from the corresponding data sensing amplifier among the plurality of data sensing amplifiers.
13. The repair circuit of claim 11, wherein the corresponding memory element comprises a corresponding column plane of the memory device.
14. The repair circuit of claim 13, wherein the repair value is received from a global column plane, the global column plane serving as a redundant storage device to be used in the repair of the failed column plane.
15. The repair circuit according to claim 11, wherein the logic gate includes an AND gate.
16. The repair circuit of claim 11, wherein each of the plurality of multiplexers is configured to receive the respective fault signal and perform selective output at least in part based on the respective fault signal.
17. The repair circuit of claim 11, further comprising an error correction circuit configured to determine, at least partially, whether data from the plurality of amplifiers contains an error while the plurality of multiplexers select the repair value of the respective output.
18. A method for memory operations, comprising: The output of the data sensing amplifier is received from the data sensing amplifier at the logic gate; The logic gate receives an indication of whether the column plane of the memory bank has been repaired. The output or default value is selectively transferred from the logic gate to the error correction circuit, at least in part, based on the instruction. The output of the data sensing amplifier is reduced to zero via the logic gate to generate the default value; The error correction circuit determines whether a bit error exists in the output. The output of the data sensing amplifier is received from the data sensing amplifier at the multiplexer; The repair value is received at the multiplexer; The instruction is received at the multiplexer; as well as Based on the instruction, the output of the data sensing amplifier or the repair value is selectively output, wherein at least a portion of the determination is performed before the selective output is completed.
19. The method of claim 18, wherein the determination of at least a portion is performed simultaneously with the selection of the output of the data sensing amplifier or the repair value.
Citation Information
Patent Citations
Semiconductor memory device and system having redundancy cells
US20130117636A1
