Peak power control in integrated memory assemblies
By enabling direct communication between the master control die and the memory die in the integrated memory assembly, power and current consumption can be adjusted in real time, overcoming the limitations of power and current management in non-volatile semiconductor memory systems and improving system performance and reliability.
Patent Information
- Application Number
- CN202110371209.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2020-06-03
- Filing Date
- 2021-04-07
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2041-04-07
AI Technical Summary
In the prior art, non-volatile semiconductor memory systems have limitations in power and current management, leading to operational failures and performance degradation. Furthermore, existing power regulation methods affect the efficiency and reliability of memory systems.
The main control die and memory die in the integrated memory assembly communicate directly via bonding pads to monitor and adjust power and current consumption in real time. Detailed power/current plans can be developed through the main control die to optimize the time periods and stages of memory operation and avoid bottlenecks in data bus communication.
It achieves efficient power and current management of the memory system, improves the performance and reliability of the memory system, avoids data bus communication pressure, and optimizes the efficiency of memory operation.
Smart Images

Figure CN113764001B_ABST
Abstract
Description
Background Art
[0001] The strong growth in demand for portable consumer electronic devices is driving the need for large-capacity storage devices. Nonvolatile semiconductor memory devices (also referred to herein as "nonvolatile memory systems"), such as flash memory storage cards, are widely used to meet the growing demand for digital information storage and exchange. Their portability, flexibility, and rugged design, along with their high reliability and large capacity have made such memory devices ideal for use in a wide variety of host electronic devices, including, for example, digital cameras, digital music players, video game consoles, PDAs, cellular phones, desktop computers, laptop computers, and notebook computers. Typically, the host electronic device provides power to the nonvolatile memory system.
[0002] A non-volatile semiconductor memory system typically contains multiple semiconductor dies ("memory dies"), each containing non-volatile memory cells that can be programmed to store data. The memory dies are typically controlled by a memory controller that is separate from the memory dies. The memory controller typically communicates with the memory dies using a standard communication protocol via a data bus. In the case of NAND memory, an open NAND flash interface (ONFI) may be used as a communication protocol. The memory dies may be operated in parallel by the memory controller in order to enhance the performance of the memory system. However, parallel operation is typically limited by the ability of the hosting device to provide power and / or current to the memory system. For example, there may be limitations on peak power and / or average power. There are typically limitations on the peak current that can be provided from the host device to the memory system. The term "peak Icc" is sometimes used to refer to the amount of peak current allowed. If the peak current drawn by the memory system is too high, the magnitude of the supply voltage may drop, which may cause operational failures in the memory system.
[0003] The relationship between current, voltage, and power is well understood by those of ordinary skill in the art. Where the voltage is substantially constant, regulation of current usage regulates power usage. For example, regulating (e.g., limiting) the peak current from a power supply having a substantially constant supply voltage also regulates (e.g., limits) the peak power. Thus, as used herein in the context of a power supply having a substantially constant voltage, the term power regulation (whether peak or average) will be understood to include current regulation. As described above, if the current drawn is too high, the voltage output of the power supply will generally not be constant. However, such a power supply will still be able to maintain a substantially constant voltage at less than the allowed peak current.
[0004] Much of the power and / or current used by a memory system is used to perform memory operations, such as programming memory cells on a memory die. Therefore, it is important to regulate the power and / or current used by the memory die in order to keep the current of the entire memory system within an allowed amount. Therefore, the amount of power and / or current that can be drawn by a memory die is typically limited.
[0005] One possible technique to control current consumption is to have circuitry in a memory die monitor the current consumption of that particular memory die. If the current consumption of a memory die exceeds a limit, the memory die temporarily suspends memory operations. A disadvantage of this technique is that a memory die may suspend its memory operations whenever other memory die in the memory system are not using much current. Therefore, it may not actually be necessary to suspend memory operations in order to keep the total current usage of the memory system within an allowed amount.
[0006] Another possible technique for controlling current consumption is for the memory controller to regulate the power and / or current consumption of the memory die. However, this solution typically requires the transmission of information over the aforementioned data bus (such as the ONFI bus). Such data transmission can potentially degrade memory system performance. In addition, it can be difficult to provide the memory controller with detailed knowledge of the power usage of the memory die. For example, a programming operation may involve multiple programming and verification loops. The number of such loops and therefore the amount of power that a complete programming operation will use is unknown before the programming operation is initiated. BRIEF DESCRIPTION OF THE DRAWINGS
[0007] Figure 1A is a block diagram of one embodiment of a memory system connected to a host.
[0008] Figure 1B is a block diagram of one embodiment of a front end processor circuit.
[0009] Figure 2 is a block diagram of one embodiment of a back-end processor circuit.
[0010] Figure 3 is a functional block diagram of an embodiment of an integrated memory assembly.
[0011] Figure 4 is a block diagram depicting one embodiment of a sensing block.
[0012] Figure 5 is a diagram showing further details of one embodiment of a circuit for applying voltages to memory structures in a memory die.
[0013] Figure 6 is a diagram of one embodiment of a processor connected to a storage device.
[0014] Figure 7 is a diagram showing various programming modules that can be stored in the memory area.
[0015] Fig. 8A is a block diagram of an embodiment of an integrated memory assembly.
[0016] Figure 8B is a block diagram of an embodiment of an integrated memory assembly in which a control die controls two memory dies.
[0017] Fig. 9A and 9B It is a top view of a semiconductor wafer.
[0018] Fig.10 Depicted is an example pattern of bond pads on a planar surface of a semiconductor die.
[0019] Fig.11A Depicted is a side view of an embodiment of an integrated memory assembly stacked on a substrate.
[0020] Fig. 11B Depicted is a side view of an embodiment of an integrated memory assembly stacked on a substrate.
[0021] Fig.12 is a perspective view of a portion of one example embodiment of a monolithic three dimensional memory array that may include a memory structure.
[0022] Fig.13A is a diagram of one embodiment of an integrated memory assembly.
[0023] Fig. 13B is a diagram of one embodiment of an integrated memory assembly in which one control die controls two memory dies.
[0024] Fig.14 is a flow chart describing one embodiment of a process for programming NAND strings of memory cells organized into an array.
[0025] Fig.15A Example threshold voltage distributions for a memory array are shown when each memory cell stores three bits of data.
[0026] Fig. 15B Depicted are threshold voltage distributions where each memory cell stores four bits of data.
[0027] Fig.16 Describe the application can be used Fig.14 Flow chart of one embodiment of a programming pulse and programming voltage of step 1408.
[0028] Fig.17 Shown in Fig.16The timing of various signals during one embodiment of a process.
[0029] Fig.18 A flow chart describing a verify operation performed to verify whether a memory cell has been programmed to its target data state is depicted.
[0030] Fig.19 A flow chart depicting one embodiment of a process for regulating power / current used by an integrated memory assembly.
[0031] Fig. 20 Depicted are several graphs showing examples of current (eg, Icc) usage profiles for several memory dies.
[0032] Fig.21 Depicted is a graph to illustrate one example of combined current usage of all memory dice in an integrated memory assembly.
[0033] Fig. 22 A flow chart depicting one embodiment of a process for master control die development scheduling.
[0034] Fig.23 Draws a graph showing the current usage of a programming loop.
[0035] Fig.24 A flow chart depicting one embodiment of a process by which a control die determines a power usage profile.
[0036] Fig.25 A flow chart depicting one embodiment of a process by which a control die performs memory operations according to a schedule provided by a master control die.
[0037] Fig.26 A flow chart depicting one embodiment of a process for a master control die to update a schedule in response to receiving notification from a slave control die that it has completed its programming operation. DETAILED DESCRIPTION
[0038] The present technology will now be described with reference to the drawings, which in embodiments relates to power and / or current regulation in an integrated memory assembly having a control semiconductor die and a memory semiconductor die. The memory semiconductor die (or "memory die") contains non-volatile memory cells. The control semiconductor die (or "control die") contains control circuits that control memory operations in the memory die. In one embodiment, the integrated memory assembly has a stack of control dies and memory dies. In some embodiments, the control die and the memory die are attached to each other by, for example, bonding pads.
[0039] In some embodiments, one of the control dies in the integrated memory assembly acts as the master control die that regulates the power and / or current usage of the integrated memory assembly. In one embodiment, each control die reports information about its expected power and / or current usage to the master control die. For example, each control die may report the expected current usage (of the memory die controlled by the corresponding control die) in each of a plurality of upcoming time periods. The master control die then determines a plan that satisfies the power and / or current criteria for the integrated memory assembly. In one embodiment, the master control die determines a plan that will maximize current usage in each time period while remaining within the current budget. The plan may include selecting which of the memory dies to perform memory operations (or phases of memory operations) during a given time period. In one embodiment, the master control die sends a die scheduling plan to each of the other control dies. Each die scheduling plan indicates when to perform memory operations and / or phases of memory operations.
[0040] Because the master control die has power / current (note that "power / current" will be used herein for average power and / or current) information about all memory dies in the integrated memory assembly, the master control die is able to develop a plan that uses the allowed power / current wisely. For example, the master control die can maximize power / current usage during each time period, which helps improve performance. For example, a memory die may be allowed to use a large amount of power / current in several time periods, provided that the total power / current usage of all memory dies does not exceed the allowed budget. If the plan actually limits the power / current of each memory die to some fixed threshold, regardless of the current usage of other memory dies, then the memory die may be shut down unnecessarily.
[0041] In addition, no information needs to be transmitted between the integrated memory assembly and the memory controller through the data bus to adjust the power and / or current use of the integrated memory assembly. In some embodiments, the memory controller communicates with the control die through the ONFI data bus. However, in some embodiments, the communication between the main control die and other control die in the integrated memory assembly does not occur through the ONFI data bus. Therefore, the communication through the ONFI data bus is not adversely affected.
[0042] In addition, the master control die is able to develop a power / current plan based on detailed information about the memory operations in the memory die. It may not be practical, or even possible, to provide such detailed information to the memory controller. For example, power / current usage may change significantly at different stages of a memory operation (such as a programming operation). A programming operation may include a programming phase followed by a verification phase (the two phases may be referred to as a programming loop). Multiple programming loops may be performed until the verification phase indicates programming memory cells. Power / current draw may be significantly different for the programming step than for the verification phase. In some embodiments, the control die reports the expected power / current information for each stage of the programming operation to the master control die, which allows fine-grained control. In addition, in some embodiments, the control die reports to the master control die when the programming operation is completed, which allows the master control die to reallocate some of the power / current budget in a specific time period to another control die. Therefore, the power / current usage in each time period can be maximized.
[0043] In some embodiments, the control die and the memory die are bonded together with many bonding pads that permit communication between the control die and the memory die. In one embodiment, the control die is able to access data from the memory die through the bonding pads. In one embodiment, each data bit and each parity bit of the codeword is read by means of different bonding pads. This actually provides an interface that is much wider than the data bus between the integrated memory assembly and the memory controller. Therefore, it is very efficient to transfer the codeword from the memory die to the control die.
[0044] In some embodiments, the control die and the memory die are manufactured on different semiconductor wafers, which permits the use of different semiconductor manufacturing processes on different wafers. For example, the semiconductor manufacturing process may involve high temperature annealing. Such high temperature annealing may be required for the proper formation of some circuit elements, but may damage other circuit elements. Due to the limitations of the semiconductor manufacturing process, it may be challenging to form complex circuits (such as decoders) on the memory die. In addition, the manufacturing process used to form memory cells on the memory die may set constraints on the size of the transistors formed on the memory die. In some embodiments, the control circuit on the control die has transistors that are different in size (e.g., smaller) than the memory cell transistors on the memory die. The different (e.g., smaller) sizes of the transistors on the control die may improve the performance of the control circuit on the control die. For example, a smaller transistor may use less power than a larger transistor. In addition, the use of smaller transistors allows one embodiment of the control die to have more transistors for the control circuit on the control die. Therefore, the control die may be better suited for operations such as determining a new read reference level than the memory die.
[0045] It should be understood that the present invention can be embodied in many different forms and should not be construed as being limited to the embodiments described herein. Specifically, these embodiments are provided so that the disclosure will be thorough and complete, and these embodiments will fully convey the present invention to those skilled in the art. In fact, the present invention is intended to cover the replacements, modifications and equivalents of these embodiments, which are included in the scope and spirit of the present invention as defined by the appended claims. In addition, in the following detailed description of the present invention, many specific details are set forth in order to provide a thorough understanding of the present invention. However, it will be appreciated by those skilled in the art that the present invention can be practiced without such specific details.
[0046] The terms "top" and "bottom", "upper" and "lower", and "vertical" and "horizontal", and forms thereof, as may be used herein, are provided for example and illustrative purposes only and are not intended to limit the description of the technology, as the items mentioned may exchange positions and orientations. Additionally, as used herein, the terms "substantially" and / or "approximately" mean that for a given application, a specified dimension or parameter may vary within acceptable manufacturing tolerances.
[0047] Figures 1A to 3 One example of a memory system that can be used to implement the techniques disclosed herein is described.
[0048] Figure 1A1 is a block diagram of one embodiment of a memory system 100 connected to a host 120. The memory system 100 may implement the techniques disclosed herein. Many different types of memory systems may be used with the techniques disclosed herein. One example memory system is a solid state drive (“SSD”); however, other types of memory systems may also be used. The memory system 100 includes a memory controller 102, an integrated memory assembly 104 for storing data, and a local memory (e.g., DRAM / ReRAM) 106. The memory controller 102 includes a front end processor circuit (FEP) 110 and one or more back end processor circuits (BEP) 112. In one embodiment, the FEP 110 circuit is implemented on an ASIC. In one embodiment, each BEP circuit 112 is implemented on a separate ASIC. The ASICs for each of the BEP circuits 112 and the FEP circuits 110 are implemented on the same semiconductor so that the memory controller 102 is manufactured as a system on a chip (“SoC”). Both the FEP 110 and the BEP 112 include their own processors. In one embodiment, the FEP 110 and the BEP 112 act as a master-slave configuration, where the FEP 110 is the master and each BEP 112 is a slave. For example, the FEP circuit 110 implements a flash translation layer that performs memory management (e.g., garbage collection, wear leveling, etc.), logical to physical address translation, communication with the host, management of DRAM (local volatile memory), and management of the overall operation of the SSD (or other non-volatile storage system). The BEP circuit 112 manages memory operations in the integrated memory assembly / die at the request of the FEP circuit 110. In some embodiments, the integrated memory assembly is referred to as a memory package. For example, the BEP circuit 112 can perform read, erase, and program processes. In addition, the BEP circuit 112 can perform buffer management, set the ratio voltage level required by the FEP circuit 110, perform error correction (ECC), control the switch mode interface with the memory package, etc. In one embodiment, each BEP circuit 112 is responsible for its own set of memory packages. The memory controller 102 is an example of a control circuit. The term device may be used herein to refer to any of, but not limited to, the integrated memory assembly 104 , the memory system 100 , the memory controller 102 , or a combination of the memory system 100 and the host 120 .
[0049] In one embodiment, there are multiple integrated memory assemblies 104. In one embodiment, each integrated memory assembly 104 includes one or more memory dies and one or more control dies. Each memory die may include one or more memory structures. The control die controls the operation of the memory die. For example, the control die may control read, write, and erase operations on the memory die. In one embodiment, the memory controller 102 communicates with the control die to instruct the control die to perform read, write, or erase operations on one or more non-volatile memory dies or one or more memory structures. In one embodiment, each memory die in the integrated memory assembly 104 utilizes NAND flash memory (including two-dimensional NAND flash memory and / or three-dimensional NAND flash memory). In other embodiments, the integrated memory assembly 104 may include other types of memory; for example, the memory package may include PCM memory.
[0050] The memory controller 102 communicates with the host 120 using an interface 130 that implements NVM Express (NVMe) over PCI Express (PCIe). To work with the memory system 100, the host 120 includes a host processor 122, a host memory 124, and a PCIe interface 126 that communicate over a bus 128. The host memory 124 is the physical memory of the host and can be DRAM, SRAM, non-volatile memory, or another type of storage device. The host 120 is external and separate from the memory system 100. In one embodiment, the memory system 100 is embedded in the host 120.
[0051] Figure 1B is a block diagram of one embodiment of the FEP circuit 110 . Figure 1BThe PCIe interface 150 is shown to communicate with the host 120 and the host processor 152 communicates with the PCIe interface. The host processor 152 can be any type of processor known in the art that is suitable for implementation. The host processor 152 communicates with a network on chip (NOC) 154. NOC is a communication subsystem on an integrated circuit, usually between cores in a SoC. NOC can span synchronous and asynchronous clock domains or use clockless asynchronous logic. NOC technology applies networking theory and methods to on-chip communication and brings significant improvements over conventional buses and crossbar interconnects. Compared to other designs, NOC improves the scalability of SoCs and the power efficiency of complex SoCs. The lines and links of the NOC are shared by multiple signals. Since all links in the NOC can operate on different data packets at the same time, a high level of parallelism is achieved. Therefore, as the complexity of the integrated subsystem continues to increase, NOC provides enhanced performance (such as throughput) and scalability compared to previous communication architectures (e.g., dedicated point-to-point signal lines, shared buses, or segmented buses with bridges). NOC 154 is connected to and communicates with memory processor 156, SRAM 160, and DRAM controller 162. DRAM controller 162 is used to operate and communicate with DRAM (e.g., DRAM 106). SRAM 160 is local RAM memory used by memory processor 156. Memory processor 156 is used to run FEP circuits and perform various memory operations. NOC also communicates with two PCIe interfaces 164 and 166. Figure 1B In the embodiment of the present invention, the memory controller 102 includes two BEP circuits 112; therefore, there are two PCIe interfaces 164 / 166. Each PCIe interface communicates with one of the BEP circuits 112. In other embodiments, there may be more or less than two BEP circuits 112; therefore, there may be more than two PCIe interfaces.
[0052] Figure 2 is a block diagram of one embodiment of the BEP circuit 112 . Figure 2 A method for communicating with the FEP circuit 110 (e.g., Figure 1B200 communicates with two NOCs 202 and 204. In one embodiment, the two NOCs can be combined into one large NOC. Each NOC (202 / 204) is connected to SRAM (230 / 260), buffer (232 / 262), processor (220 / 250) and data path controller (222 / 252) via an XOR engine (224 / 254), an ECC engine (226 / 256) and a read reference voltage calibration engine (225 / 255). The ECC engine 226 / 256 is used to perform error correction, as known in the art. Herein, the ECC engine 226 / 256 may be referred to as a controller ECC engine. The XOR engine 224 / 254 is used to XOR the data so that the data can be combined and stored in a manner that can be recovered in the presence of programming errors.
[0053] The data path controller 222 is connected to the memory interface 228 to communicate with the integrated memory assembly via four channels. Therefore, the top NOC 202 is associated with the memory interface 228 of the four channels for communicating with the integrated memory assembly, and the bottom NOC 204 is associated with the memory interface 258 of the four additional channels for communicating with the integrated memory assembly. In one embodiment, each memory interface 228 / 258 includes four switching mode interfaces (TM interfaces), four buffers and four schedulers. There is a scheduler, buffer and TM interface for each of the channels. The processor can be any standard processor known in the art. The data path controller 222 / 252 can be a processor, FPGA, microprocessor or other type of controller. The XOR engine 224 / 254, the ECC engine 226 / 256 and the read reference voltage calibration engine (225 / 255) are dedicated hardware circuits, which are referred to as hardware accelerators. In other embodiments, the XOR engine 224 / 254 and the ECC engine 226 / 256 can be implemented in software. The scheduler, buffer, and TM interface are hardware circuits. In other embodiments, the memory interface (circuitry for communicating with the memory die) can be a Figure 2 In addition, there is a different structure than that depicted in Figure 1B and 2 Controllers of different configurations may also be used with the techniques described herein.
[0054] Figure 3304. The integrated memory assembly 104 includes two types of semiconductor dies (or more succinctly, "dies"). One type is referred to herein as a control die 304 and the other type is referred to herein as a memory die 302. The memory die 302 includes a memory structure 326. The memory structure 326 may contain non-volatile memory cells. The control die 304 includes a control circuit 310. In some embodiments, the memory die 302 and the control die 304 are bonded together, as will be described in more detail below. Typically, there are multiple control dies 304 and multiple memory dies 302 in the integrated memory assembly 104.
[0055] The control circuit 310 performs memory operations (e.g., write, read, erase, etc.) on the memory structure 326. The control circuit 310 includes a processor 312, an on-chip address decoder 314, a power control circuit 316, a storage area 318, a read / write circuit 328, a command interpreter 330, and a memory controller interface 334. In another embodiment, a portion of the read / write circuit 328 is located on the control die 304, and a portion of the read / write circuit 328 is located on the memory die 302. For example, the read / write circuit 328 may contain a sense amplifier. In one embodiment, the sense amplifier is located on the control die 304. In one embodiment, the sense amplifier is located on the memory die 302.
[0056] As used herein, the term "memory die," "memory semiconductor die," or similar terms means a semiconductor die containing non-volatile memory cells for storage. As used herein, the term "control die," "control semiconductor die," or similar terms means a semiconductor die containing control circuits for performing memory operations on non-volatile memory cells on a memory die. Typically, many semiconductor dies 304 are formed from a single semiconductor (e.g., silicon) wafer. Typically, many memory dies 302 are formed from a single semiconductor (e.g., silicon) wafer.
[0057] Processor 312 is programmable and reprogrammable. Processor 312 can be a microprocessor. In one embodiment, the processor has a reduced instruction set computer (RISC) architecture. Processor 312 executes programming instructions to control memory die 302. Processor 312 controls the operation of address decoder 314, power control 316 and read / write circuit 328. Therefore, processor 312 is capable of controlling memory operations in memory structure 326. In some embodiments, processor 312 executes instructions to adjust the power and / or current use of integrated memory assembly 104. In some embodiments, processor 312 is replaced with a hard-wired state machine. Therefore, the hard-wired state machine can be used to control the operation of address decoder 314, power control 316 and read / write circuit 328, and thus control the memory operations in memory structure 326. In some embodiments, control die 304 has both a hard-wired state machine and a processor (e.g., a microprocessor).
[0058] The memory area 318 may be used to store programming instructions executed on the processor 312. The memory area 318 may be used to store various parameters and settings for controlling the memory die 302. For example, the parameters and settings may include voltage levels of signals applied to the memory structure 326. The memory area 318 may contain volatile memory (e.g., SRAM, DRAM) or non-volatile memory.
[0059] The memory structure 326 also has a memory structure storage area 336 to store programming instructions executed on the processor 312. The memory structure storage area 336 can be used to store various parameters and settings for controlling the memory die 302. In some embodiments, when the integrated memory assembly 104 is initialized (e.g., booted up), the control die 304 transfers the programming instructions, parameters, settings, and the like from the memory structure storage area 336 to the storage area 318. In an embodiment, the control die 304 has hardware (e.g., a hardwired state machine) to control the transfer from the memory structure storage area 336 to the storage area 318. The memory structure storage area 336 can be implemented in a non-volatile memory unit that is not used to store user data.
[0060] In some embodiments, programming instructions executed by processor 312 are updated by replacing a portion of the programming instructions with updated (or modified) instructions. Various parameters and / or settings may also be updated. Such updates allow changes to be made to how processor 312 operates memory die 302. Such updates (or changes) may not be possible or practical to make in hardware (e.g., a hardwired state machine).
[0061] In some embodiments, each memory structure 326 can be decoded by a row decoder ( Figure 3The on-chip address decoder 314 provides an address interface between the address used by the host 120 or the memory controller 102 to the row decoder and the column decoder (not depicted). Figure 3 The power control circuit 316 controls the power and voltage supplied to the word lines, bit lines, and select lines during memory operations. The power control circuit 316 may include a voltage generator (e.g., a charge pump) for generating voltages. The power control circuit 316 executes under the control of the processor 312.
[0062] In some embodiments, read / write circuit 328 includes a sense block (which may contain a sense amplifier (SA)). In some embodiments, the sense amplifier includes a bit line driver. Read / write circuit 328 executes under the control of processor 312.
[0063] The command interpreter 330 is capable of interpreting commands provided by the memory controller 102. For example, the command interpreter 330 can interpret read, write, erase, and other commands. In some embodiments, the command interpreter 330 is implemented in hardware (such as a state machine). Optionally, the command interpreter 330 can be implemented by software executed on the processor 312.
[0064] The memory controller interface 334 is an electrical interface for communicating with the memory controller 102. For example, the memory controller interface 334 may implement a switch mode interface that is connected to a switch mode interface of the memory interface 228 / 258 for the memory controller 102. In one embodiment, the memory controller interface 334 includes a collection of input and / or output (I / O) pins connected to a communication channel 332 (also referred to herein as a data bus). In one embodiment, the communication channel 332 is connected to the memory controller 102 as part of the switch mode interface. In one embodiment, the communication channel 332 of one integrated memory assembly 104 is connected to another integrated memory assembly 104.
[0065] The communication channel 332 is depicted as being connected to the integrated memory assembly 104 for generality. The communication channel 332 may be connected to either or both of the dies 302 and / or 304. In one embodiment, the communication channel 332 connects the memory controller 102 directly to the control die 304. In one embodiment, the communication channel 332 connects the memory controller 102 directly to the memory die 302. If the communication channel 332 connects the memory controller 102 directly to the memory die 302, then the path 352 may be used to allow communication between the memory controller 102 and the control circuit 310.
[0066] Any subset of the components in control circuit 310 may be considered to be one or more control circuits. Figure 3 The processor 312, command interpreter 330, read / write circuit 328, and / or memory controller 102 (or circuits acting equivalently) in combination with all other circuits depicted in or a subset of other circuits may be considered to be one or more control circuits. The one or more control circuits may include hardware alone or a combination of hardware and software (including firmware). For example, a controller programmed by firmware is an example of a control circuit. The one or more control circuits may include a processor, a programmable gate array (PGA), a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), an integrated circuit, or other type of circuit.
[0067] Path 352 is a path between one or more components in control circuit 310 and a memory structure on memory die 302. A portion of each path resides in memory die 302 and a portion of each path resides in control die 304. The term path can be used for a portion of path 352 that is completely within one of the dies. Therefore, it can be said that memory die 302 has a first plurality of paths and control die 304 has a second plurality of paths. In one embodiment, control die 304 and memory die 302 are configured to transmit signals through a path pair of the first plurality of paths and the second plurality of paths. In some embodiments, memory die 302 and control die 304 are bonded to each other, or otherwise attached to each other, to facilitate signal transmission through the path pair.
[0068] Paths can be used to provide or receive signals (e.g., voltage, current). Paths include conductive paths. Paths may include one or more of the following (but not limited to): bonding pads, metal interconnects, vias, transistors, conductive materials, and other materials that can transmit or carry electrical signals. In one embodiment, paths 352 allow control circuit 310 to provide voltage to word lines, select lines, and bit lines on memory die 302. Paths 352 can be used to receive signals from, for example, bit lines. In one embodiment, there are about 100,000 paths 352. However, there may be more or less than 100,000 paths. Having such a large number of paths 352 allows a very large amount of data or other signals to be transmitted in parallel.
[0069] In one embodiment, the memory structure 326 includes a three-dimensional memory array of non-volatile memory cells with multiple memory levels formed over a single substrate such as a wafer. The memory structure may include any type of non-volatile memory monolithically formed in one or more physical levels of an array of memory cells having active regions disposed over a silicon (or other type of) substrate. In one example, the non-volatile memory cells include vertical NAND strings having charge trapping material.
[0070] In another embodiment, the memory structure 326 includes a two-dimensional memory array of non-volatile memory cells. In one example, the non-volatile memory cells are NAND flash memory cells that utilize floating gates. Other types of memory cells (eg, NOR-type flash memory) may also be used.
[0071] The exact type of memory array architecture or memory cells included in the memory structure 326 is not limited to the above examples. Many different types of memory array architectures or memory technologies can be used to form the memory structure 326. For the purposes of the newly claimed embodiments presented herein, no specific non-volatile memory technology is required. Other examples of suitable technologies for the memory cells of the memory structure 326 include phase change memory (e.g., PCM) and the like. Examples of suitable technologies for the memory cell architecture of the memory structure 326 include two-dimensional arrays, three-dimensional arrays, cross-point arrays, stacked two-dimensional arrays, vertical bit line arrays, and the like.
[0072] Those of ordinary skill in the art will recognize that the technology described herein is not limited to a single particular memory structure, but encompasses many related memory structures within the spirit and scope of the technology as described herein and as understood by those of ordinary skill in the art.
[0073] Although Figure 3 One control die 304 and one memory die 302 are depicted in the integrated memory assembly 104, but there may be more than one control die 304 and more than one memory die 302 in the integrated memory assembly 104. In some cases, a single control die 304 may control more than one memory die 302.
[0074] Typically, there are multiple control dies 304 and multiple memory dies 302 in the integrated memory assembly 104. In some embodiments, one of the control dies 304 is a master control die. In some embodiments, the master control die determines a power plan for operating memory operations in the integrated memory assembly 104. The processor 312 can be used to determine the power plan based on future memory operations (and / or phases of memory operations) to be performed in the memory dies 302 in the integrated memory assembly 104. Each control die 304 in the integrated memory assembly 104 can send the master control die information about future power consumption so that the master control die can determine the power plan. In some embodiments, the master control die determines which memory dies 304 perform memory operations in each of several upcoming time periods. In one embodiment, the master control die sends a schedule to each other control die, the schedule indicating when the control die should perform the memory operations (or phases thereof) in the memory die it controls.
[0075] Figure 4 4 is a block diagram depicting one embodiment of a sense block 450 operating under the control of a processor 312. The sense block is part of the read / write circuit 328. There may be many sense blocks 450 in the read / write circuit 328. The individual sense blocks 450 are divided into one or more core portions called sense circuits or sense amplifiers 350(1) to 350(4) and a common portion called management circuit 360. In one embodiment, there will be a separate sense amplifier for each bit line / NAND string and one common management circuit 360 for a set of multiple (e.g., four or eight) sense amplifiers. Each of the sense amplifiers in the group communicates with the associated management circuit 360 via a data bus 454. Thus, there are one or more latch management circuits that communicate with the sense amplifiers of the set of storage elements (memory cells). The processor 312 controls the overall operation of the sense amplifiers 350 and the management circuit 360; however, some aspects of the sense amplifiers 350 and the management circuit 360 may operate independently of the control of the processor 312.
[0076] As an example, sense amplifier 350(1) includes sense circuitry 460 that senses by determining whether the conduction current in a connected bit line is above or below a predetermined threshold level. Sensing may occur in a read or verify operation. The sense amplifier also supplies the bit line voltage during application of a programming voltage in a programming operation. Processor 312 controls the timing and magnitude of the bit line voltage.
[0077] The sensing circuit 460 may include a Vbl selector 462, a sensing node 464, a comparison circuit 466, and a trip latch 468. During the application of the programming voltage, the Vbl selector 462 may pass a programming enable voltage (e.g., V_pgm_enable) or a programming inhibit voltage (e.g., Vbl_inh) to the bit line connected to the memory cell. Herein, a "programming enable voltage" is defined as a voltage applied to a memory cell that enables programming of the memory cell while a programming voltage (e.g., Vpgm) is also applied to the memory cell. In some embodiments, the programming enable voltage is applied to the bit line coupled to the memory cell while a programming voltage is applied to the control gate of the memory cell. Herein, a "programming inhibit voltage" is defined as a voltage applied to a bit line coupled to the memory cell to inhibit programming of the memory cell while a programming voltage (e.g., Vpgm) is also applied to the memory cell (e.g., applied to the control gate of the memory cell). It should be noted that a boost voltage (e.g., Vpass) may be applied to an unselected word line and a programming inhibit voltage is applied to a bit line.
[0078] The program inhibit voltage is applied to the bit lines coupled to the memory cells that are not programmed and / or the bit lines with the memory cells that have reached their corresponding target threshold voltages by executing the programming process. These may be referred to as "unselected bit lines". The program inhibit voltage is not applied to the bit lines with the memory cells to be programmed ("selected bit lines"). In one embodiment, when the program inhibit voltage is applied to the unselected bit lines, the bit lines are cut off from the NAND channel. Therefore, in one embodiment, the program inhibit voltage is not passed to the NAND channel. The boost voltage is applied to the unselected word lines to raise the potential of the NAND channel, which inhibits the programming of the memory cells that receive the programming voltage at their control gates.
[0079] Transistor 470 (e.g., nMOS) can be configured as a pass-gate to pass Vbl from Vbl selector 462 by setting the control gate voltage of the transistor high enough (e.g., higher than Vbl passed from the Vbl selector). For example, selector 472 can pass the power supply voltage Vdd (e.g., 3 to 4V) to the control gate of transistor 470.
[0080] Sense amplifier 350(1) is configured to control the timing of when voltage is applied to the bit line under the control of processor 312. During sensing operations (such as read and verify operations), the bit line voltage is set by transistor 470 based on the voltage passed by selector 472. The bit line voltage is approximately equal to the control gate voltage of the transistor minus its Vt (e.g., 3V). For example, if Vbl+Vt is passed by selector 472, then the bit line voltage will be Vbl. This assumes that the source line is at 0V. Transistor 470 clamps the bit line voltage based on the control gate voltage and acts as a source follower rather than a pass gate. Vbl selector 462 can pass a relatively high voltage (such as Vdd) that is higher than the control gate voltage on transistor 470 to provide a source follower mode. During sensing, transistor 470 thereby charges the bit line.
[0081] In one approach, the selector 472 of each sense amplifier can be controlled separately from the selectors of the other sense amplifiers to pass either Vbl or Vdd. The Vbl selector 462 of each sense amplifier can also be controlled separately from the Vbl selectors of the other sense amplifiers.
[0082] During sensing, the sense node 464 is charged to an initial voltage, such as Vsense_init=3V. The sense node is then connected to the bit line via transistor 470, and the amount of decay of the sense node is used to determine whether the memory cell is in a conductive state or a non-conductive state. In one embodiment, a current flowing in a bot line discharges the sense node (e.g., a sense capacitor). The length of time the sense node is allowed to decay may be referred to herein as an "integration time." A comparison circuit 466 is used to compare the sense node voltage with a trip voltage at the sensing time. If the sense node voltage decays below the trip voltage Vtrip, then the memory cell is in a conductive state and its Vt is at or below the voltage of the verification signal. If the sense node voltage does not decay below Vtrip, then the memory cell is in a non-conductive state and its Vt is above the voltage of the verification signal. The sense amplifier 350(1) includes a trip latch 468 that is set by the comparison circuit 466 based on whether the memory cell is in a conductive or non-conductive state. The data in the trip latch may be a bit read out by a latch manager 482. In some embodiments, processor 312 performs calculations to determine the data stored in the sensed memory cell. To perform such calculations, latch manager 482 may provide processor 312 with data (eg, bits) from trip latch 468.
[0083] The management circuit 360 includes a latch manager 482, four example sets of data latches 484, 485, 486, 487, and an I / O interface 488 coupled between the set of data latches and the data bus 332 (the data bus may be connected to the memory controller 102). A set of data latches (e.g., including individual latches LDL, LMDL, UMDL, and UDL) may be provided for each sense amplifier. In some cases, fewer or additional data latches may be used. The LDL stores bits for the lower page of data, the LMDL stores bits for the lower middle page of data, the UMDL stores bits for the upper middle page of data, and the UDL stores bits for the upper page of data. This is a sixteen-level or four-bit per memory cell memory device. In one embodiment, there are eight levels or three bits per memory cell.
[0084] The latch manager 482 operates under the control of the processor 312. However, the latch manager 482 may perform some operations independently of the control of the processor 312. In some embodiments, the latch manager 482 is implemented in hardware, which may include but is not limited to discrete logic (e.g., XOR gates, AND gates, inverters, etc.). In some embodiments, the latch manager 482 performs calculations to determine the data stored in the sensed memory cell. In one embodiment, to determine the data stored in the sensed memory cell, the latch manager 482 may contain discrete logic, such as but not limited to XOR gates. The latch manager 482 may also contain latches that can be used to store information to help determine the data stored in the sensed memory cell. However, using discrete logic to determine the data stored in the sensed memory cell may have limitations, such as not being able to change the page mapping scheme. In some embodiments, using the processor 312 to determine the data in the memory cell allows the page mapping scheme to be changed. Other details of embodiments of changing the page mapping scheme are described below.
[0085] The determined data (whether determined by processor 312 or independently by latch manager 482) can be stored in a set of data latches. Each set of data latches 484-487 is used to store data bits determined during a read operation and to store data bits imported from data bus 332 during a program operation, the data bits representing write data to be programmed into the memory. I / O interface 488 provides an interface between data latches 484-487 and data bus 332.
[0086] In one embodiment, latch manager 482 is used to determine what voltage to apply to the bit line based on the state of the latch. Thus, latch manager 482 can perform some operations independently of processor 312, thereby alleviating the need for processor 312 to perform all calculations and / or decisions.
[0087] During read, the operation of the system is under the control of the processor 312 which controls the supply of different control gate voltages to the addressed memory cells. As it steps through the various predefined control gate voltages corresponding to the various memory states supported by the memory, the sense circuit 460 may trip at one of these voltages and the corresponding output will be provided from the sense circuit 460 to the latch manager 482 via the data bus 454. In one embodiment, the latch manager 482 determines the resulting memory state by taking into account the trip event of the sense circuit 460 and the information about the control gate voltage applied from the processor 312 via the input line 490. The latch manager 482 then calculates the binary encoding of the memory state and stores the resulting data bits into the data latches 484-487.
[0088] During a program or verify operation for a memory cell, the data to be programmed (write data) is stored in a data set of latches 484-487 from data bus 332 in the LDL, LMDL, UMDL and UDL latches in a four-bit per memory cell implementation.
[0089] A programming operation under the control of processor 312 applies a set of programming voltage pulses to the control gates of the addressed memory cells. Each voltage pulse may be a step increase in magnitude from the previous programming pulse in a process known as incremental step pulse programming. Each programming voltage is followed by a verify operation to determine whether the memory cell has been programmed to the desired memory state. In some cases, latch manager 482 monitors the read back memory state relative to the desired memory state. When the two agree, latch manager 482 sets the bit line in a program inhibit mode, such as by updating its latch. This inhibits further programming of the memory cell coupled to the bit line, even if additional programming pulses are applied to its control gate.
[0090] Each set of data latches 484 to 487 may be implemented as a stack of data latches for each sense amplifier. In one embodiment, there are three data latches for each sense amplifier 350. In some embodiments, the data latches are implemented as shift registers so that parallel data stored therein is converted into serial data for the data bus 332, and vice versa. All data latches corresponding to a read / write memory cell block may be linked together to form a block shift register so that a block of data may be input or output by serial transmission. Specifically, the read / write circuit group is adapted so that each of its data latch sets will sequentially shift data into or out of the data bus, as if it were part of a shift register for the entire read / write block.
[0091] The data latches identify when the associated memory cell has reached a particular milestone in a programming operation. For example, the latches may identify that the Vt of a memory cell is below a particular verify voltage. The data latches indicate whether the memory cell is currently storing one or more bits from a page of data. For example, an LDL latch may be used to store a lower page of data. When a lower page bit is stored in the associated memory cell, the LDL latch is flipped (e.g., from 0 to 1). When a lower middle, upper middle, or upper page bit, respectively, is stored in the associated memory cell, the LMDL, UMDL, or UDL latch is flipped. This occurs when the associated memory cell has completed programming.
[0092] Figure 5 302 is a diagram showing further details of one embodiment of a circuit for applying voltage to a memory structure 326 in a memory die 302. Two blocks 500, 520 of memory cells are depicted. These are two of many blocks in the memory structure 326. In one embodiment, Figure 5 The remainder of the circuitry in resides on the control die 304. However, some of the other circuitry may reside on the memory die 302. In one embodiment, the page buffer 540 resides on the memory die 302. In one embodiment, both the page buffer 540 and the column decoder 132 reside on the memory die 302.
[0093] Figure 5 The circuits in operate under the control of processor 312. Figure 5The processor 312 may include a circuit that applies voltages to the word lines, select lines, and bit lines. For example, the processor 312 may issue control signals to one or more of the CONTROL 580, the high voltage (HV) generator (GEN) 572, the page buffer (PB) driver (DRV) 574, the command (CMD) register (REG) 576, and the input / output (I / O) buffer 578. In one embodiment, the processor 312 issues control signals to the CONTROL 580, which in turn controls other components such as the HV GEN 572 and the PB DRV 574.
[0094] A high voltage generator (HV GEN) 572 is connected to the word line driver 550 and to a page buffer driver (PBDRV) 574. In one embodiment, the word line driver 550 is controlled by the processor 312 to control the magnitude and timing of the voltages of the word lines and select lines in the memory structure 326. The PB DRV 574 is connected to the page buffer 540 to control the page buffer 540. The page buffer 540 may include a sense block, such as Figure 4 In one embodiment, processor 312 controls PB DRV 574 to control the magnitude and timing of the voltage applied to bit line (BL) 542 .
[0095] In one possible embodiment, each block of storage elements is associated with a set of transfer transistors. For example, block 500, which in this example is a selected block (e.g., a block in which a programming or sensing operation occurs), includes a drain-side select gate (SGD) connected to transfer transistor 504, a drain-side dummy word line (WLDD) connected to transfer transistor 506, a word line (WL47) connected to transfer transistor 508, intermediate word lines WL46 to WL1 (not shown) connected to respective transfer transistors (not shown), a word line (WL0) connected to transfer transistor 510, a source-side dummy word line (WLDS) connected to transfer transistor 512, and a source-side select gate (SGS) connected to transfer transistor 514. The control gate of each transfer transistor of block 500 is connected to a block decoder (BD) 502 via a common path 503. BD 502 receives a voltage from a transfer transistor driver (TTDRV) 564 and a control signal from an address register (ADD REG) 570. The control signal includes an address. If the address matches the address of BD 502, then BD 502 acts as a conductive switch that passes a voltage to the control gate of the associated pass transistor via path 503. If the address does not match the address of BD 502, then BD 502 acts as a non-conductive switch that does not pass a voltage to the control gate of the associated pass transistor.
[0096] Each transfer transistor may be an nMOSFET, for example, having a drain node on the left side and a source node on the right side. The drain node for each transfer transistor is connected to a corresponding voltage driver in the set of high voltage voltage drivers 550. Each driver may be provided with a voltage from HV GEN 572. For example, transfer transistor 504 is connected to a drain select gate driver (SGD DRV) 552, transfer transistor 506 is connected to a dummy word line driver (WLDD DRV) 554, transfer transistor 508 is connected to a word line driver (WL47 DRV) 556, transfer transistor 510 is connected to a word line driver (WL0DRV) 558, transfer transistor 512 is connected to a source side dummy word line driver (WLDS DRV) 560, and transfer transistor 514 is connected to a source select gate driver (SGS DRV). Each voltage driver may be independently controlled by processor 312 to provide a desired output voltage.
[0097] A similar arrangement is used for the example unselected block 520, which includes a pass transistor 524 connected to SGD and SGD DRV 552, a pass transistor 526 connected to WLDD and WLDD DRV 554, a pass transistor 528 connected to WL47 and WL47 DRV 556, ..., a pass transistor 530 connected to WL0 and WL0 DRV 558, a pass transistor 532 connected to WLDS and WLDS DRV 560, and a pass transistor 534 connected to SGS and SGS DRV 562. The control gates of the pass transistors of the unselected block 520 are connected to the corresponding block decoder (BD) 522 via a common path 523. The BD 522 is also connected to the TT DRV 564 to receive a voltage, and to the address register 570 to receive a control signal that instructs the BD 522 to pass or not pass a voltage to the control gate of the associated pass transistor via the path 523. An address register (ADD REG) 570 also communicates with the voltage drivers in the set of high voltage drivers 550 .
[0098] In response to the column decoder 132, a number of bit lines (BL) 542 extend across the selected block 500 and the unselected block 520 (and in some embodiments, across additional unselected blocks, not shown) and extend into the page buffer 540. The page buffer stores data written to or read from the selected word line of the selected block. During operation of the memory device, the address register 570 provides data load commands to the input-output buffer 578 and the command register 576. The input-output buffer 578 provides commands to the page buffer 540. The command register 576 provides commands to the control circuit 580, which instructs the high voltage generator (HV GEN) 572 to control the voltage driver 550 at the appropriate level. Typically, during programming, the driver of the selected word line provides a programming level voltage, such as 12 to 26V, while the unselected word lines receive a pass voltage V PASS , such as 4 to 6V. During sensing, the driver of the selected word line provides a read or verify level voltage (respectively, V CGR or V VERIFY ) while the unselected word lines receive the read pass voltage, V READ-PASS Control 580 also instructs page buffer driver (PB DRV) 574 to control page buffer 540. Address register 570 also communicates with column decoder 132.
[0099] Figure 6 is a diagram depicting one embodiment of a memory system 100 connected to a host 120. The memory system 100 has a memory controller 102 and an integrated memory assembly 104. The memory controller 102 is connected to the integrated memory assembly 104 via a data bus 332. The data bus 332 is connected to a memory controller interface 334. As described above, the memory controller interface 334 may implement a switching mode interface that is connected to a switching mode interface of a memory interface 228 / 258 for the memory controller 102. In one embodiment, the data bus 332 includes an ONFI bus. The memory controller 102 sends memory commands (e.g., read, write, erase) to each control die 304 via the data bus 332. Each control die 304 performs memory operations in the memory die 302 that it is responsible for controlling. Figure 6 , each control die 304 is depicted as being connected to (and controlling) one memory die 302. However, a control die 304 can control more than one memory die 302.
[0100] The integrated memory assembly 104 has a master control die 304a, and several slave control die 304b, ... 304n. For ease of discussion, reference 304b will be used herein to refer to the slave control die, it should be understood that any control die 304 other than the master control die 304a can be a slave control die. The host 120 has a power supply 600, which provides power to the memory controller 102 and the integrated memory assembly 104. In some embodiments, the power supply 600 is capable of supplying a substantially constant voltage as long as the current is below the peak current. The master control die 304a manages the power / current consumption in the integrated memory assembly 104.
[0101] The die scheduler logic 602 determines the memory die schedule in order to implement the power plan for the integrated memory assembly 104. The memory die schedule specifies when each memory die 302 performs a memory operation (or its phase). The master control die 304a sends the memory die schedule to each slave control die 304b so that the slave control die 304b knows when it performs a memory operation (or its phase). Each control die 304 has access to a common clock signal (CLK) that is used to synchronize memory operations. That is, the common clock is used to determine when (e.g., in what time period) a memory operation (or its phase) is performed.
[0102] In one embodiment, each control die 304 has power profile generation logic 604 that determines the upcoming power (or current) usage of the memory die 302. This power / current usage is based on the upcoming memory operations to be performed in the memory die 302, and the expected power / current consumption of those operations. There may be some variation between the memory dies 302 in their power / current consumption for a particular memory operation. Therefore, in some embodiments, the control die 304 will test the memory die 302 it controls to determine the power / current characteristics that are unique to the memory die 302. However, such testing need not be performed by the control die 304. Alternatively, the memory controller 102 may determine the power / current characteristics that are unique to each memory die 302. Examples of power profiles for various memory dies 302 are depicted in Fig. 20 middle.
[0103] The master control die 304a has power profile analysis logic 606 that analyzes power profiles from various memory dies 302 (and memory dies 302 controlled by the master control die). Memory die scheduling is determined based on the power profile analysis. It should be noted that in some embodiments, the master control die 304a determines the memory die scheduling based on the expected current usage of the memory operations performed in the assembly 104.
[0104] In some embodiments, the master control die 304a communicates with the slave control die 304b, ... 304n via a communication link 608. In some embodiments, the control die 304 has through silicon vias (TSVs), and at least a portion of the communication link 608 runs through the TSVs. Communication over the communication link 608 avoids using the data bus 332. Therefore, power management does not burden the data bus 332.
[0105] In some embodiments, the master control die 304a has a processor (e.g., a microprocessor) that executes instructions to implement the die scheduler logic 602, the power profile generation 604, and the profile analysis logic 606. The slave control die 304 may also have a processor to implement the power profile generation logic 604. However, it is not required that all of the aforementioned logic (e.g., 602, 604, 606) be implemented by software (i.e., by executing instructions on a processor). In various embodiments, with respect to Figure 6 Some or all of the aforementioned logic described (e.g., 602, 604, 606) is implemented in hardware, which may include, but is not limited to: hardwired state machines, circuits, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), digital logic circuits, analog circuits, gates, flip-flops, latches, RAM / ROM, and / or combinational logic.
[0106] As mentioned, control die 304 may have a processor to perform various functionality described herein. Figure 7 3 is a diagram of one embodiment of a processor 312 on a control die 304. The processor 312 has a control unit 710, an arithmetic logic unit (ALU), and input / output (I / O) 730. Figure 7 In the embodiment of the present invention, the storage device 318 is divided between the instruction storage device 318a and the data storage device 318b. There is a path 740 between the control unit 710 and the instruction storage device 318a, and there is another path 750 between the control unit 710 and the data storage device 318b. However, there need not be separate instruction storage devices 318a and data storage devices 318b. In addition, there need not be two paths 740, 750. In one embodiment, there is a single path between the control unit 710 and the storage device 318.
[0107] In some embodiments, the control unit 710 contains an instruction register and a programming counter. The programming counter stores the address of the programming instruction in the instruction storage device 318a. The control unit 710 may also contain an instruction extractor configured to extract the instruction identified by the programming counter from the instruction storage device 318a. The control unit 710 may also contain an instruction decoder configured to decode the instruction. The control unit 710 may also contain an instruction executor configured to execute the decoded instruction. The control unit 710 may also contain registers that can be used for temporary storage of parameters.
[0108] ALU 720 is used to perform arithmetic operations, such as determining the magnitude of a voltage applied to memory structure 326. In one embodiment, the ALU performs arithmetic and bit-by-bit operations on integer binary numbers. In one embodiment, ALU 720 is a combinational digital circuit. For example, ALU 720 may contain discrete logic, such as an AND gate, a NAND gate, a NOR gate, an OR gate, an XOR gate, etc. In some embodiments, ALU 720 may be referred to as a processing unit.
[0109] In one embodiment, input / output (I / O) 730 receives commands from command interpreter 330. For example, command interpreter 330 may send read, write (or program), erase, or other commands to I / O 730. Control unit 710 accesses and executes appropriate instructions from instruction storage device 318a for the command. Control unit 710 may also access various parameters and settings from data storage device 318b. Control unit 710 instructs I / O 730 to issue control signals in order to control various circuits to affect memory operations. I / O 730 sends control signals to circuits such as power control 316 and read / write circuit 328.
[0110] Processor 312 may have a wide variety of instruction set architectures. In some embodiments, processor 312 has a RISC architecture. In one embodiment, processor 312 has a RISC-V architecture. However, processor 312 is not limited to a RISC architecture.
[0111] Processor 312 is programmable and reprogrammable. Processor 312 can be programmed by storing instructions in instruction storage device 318a, which are then executed in processor 312. Processor 312 can be reprogrammed by updating or otherwise modifying instructions in instruction storage device 318a, which are then executed in processor 312. In some embodiments, when control die 304 starts up, programming instructions are loaded from memory structure storage area 336 to instruction storage device 318a. Parameters and / or settings can be loaded from memory structure storage area 336 to data storage device 318b.
[0112] Elements in processor 312 may include, but are not limited to, one or more of a microprocessor, circuit, application specific integrated circuit (ASIC), field programmable gate array (FPGA), digital logic circuit, analog circuit, gate, flip-flop, latch, RAM / ROM, and / or combinational logic.
[0113] Fig. 8A is a block diagram of one embodiment of an integrated memory assembly 104 . Fig. 8A Depiction Figure 1A 3 or 3. Memory die 302 contains a plane 820 of memory cells. Memory die 302 may have additional planes. The plane is divided into M blocks. In one example, each plane has about 1040 blocks. However, a different number of blocks may also be used. In one embodiment, a block including memory cells is a unit of erasure. That is, all memory cells of a block are erased together. In other embodiments, memory cells may be grouped into blocks for other reasons in order to organize memory structure 326 to enable signaling and selection circuits. A representative bit line (BL) is depicted for each plane. There may be thousands or tens of thousands of such bit lines per plane. Each block may be divided into several word lines, as will be described more fully below. In one embodiment, a block represents a group of connected memory cells because the memory cells of a block share a common set of uninterrupted word lines and uninterrupted bit lines. In the structure of FIG. 8A , block 0 and block M−1 of plane 820 are at the edge of the memory structure (or otherwise referred to as being located in an edge region / portion of the memory structure).
[0114] In one embodiment, the control die 304 includes several sense amplifiers (SA) 350. In an alternative embodiment, the sense amplifiers (SA) 350 reside on the memory die 302. The processor 312 controls the sense amplifiers 350 (whether on the control die 304 or the memory die 302). The operation of the sense amplifiers 350 can be modified by updating the instructions and / or parameters used by the processor 312. In this example, each sense amplifier 350 is connected to a bit line. In one embodiment, the sense amplifier contains a bit line driver. Therefore, the sense amplifier can provide a voltage to the bit line to which it is connected. The sense amplifier is configured to sense the condition of the bit line. In one embodiment, the sense amplifier is configured to sense the current flowing in the bit line. In one embodiment, the sense amplifier is configured to sense the voltage on the bit line.
[0115] The control die 304 includes a number of word line drivers 860 (1) to 860 (n). The word line driver 860 is configured to provide voltage to the word line. In this example, there are "n" word lines per memory cell block. In one embodiment, one of the blocks in the plane 820 is selected at a time for memory array operation. In one embodiment, if the memory operation is programming or reading, then one word line within the selected block is selected for the memory operation. In one embodiment, if the memory operation is erasing, then all word lines within the selected block are selected for erasing. The word line driver 860 provides voltage to the word line in the first selected block (e.g., block 2) in the memory die 302. The control die 304 may also include a charge pump, a voltage generator, and the like that can be used to provide voltage for the word line driver 860 and / or the bit line driver.
[0116] Memory die 302 has a number of bonding pads 870a, 870b on a first major surface 882 of memory die 302. There may be "n" bonding pads 870a to receive voltages from corresponding "n" word line drivers 860(1) through 860(n). There may be one bonding pad 870b for each bit line associated with plane 820. Reference numeral 870 will be used to generally refer to bonding pads on major surface 882.
[0117] In some embodiments, each data bit and each parity bit of the codeword are transmitted through different bonding pad pairs 870b, 874b. The bits of the codeword can be transmitted in parallel through bonding pad pairs 870b, 874b. This provides extremely efficient data transmission relative to, for example, transmitting data between the memory controller 102 and the integrated memory assembly 104. For example, the data bus between the memory controller 102 and the integrated memory assembly 104 can, for example, provide eight, sixteen or possibly 32 bits transmitted in parallel. However, the data bus between the memory controller 102 and the integrated memory assembly 104 is not limited to these examples.
[0118] The control die 304 has a plurality of bonding pads 874a, 874b on a first major surface 884 of the control die 304. There may be "n" bonding pads 874a to deliver voltages from corresponding "n" word line drivers 860(1) to 860(n) to the memory die 302a. There may be one bonding pad 874b for each bit line associated with the plane 820. Reference numeral 874 will be used to generally refer to the bonding pads on the major surface 882. It should be noted that there may be bonding pad pairs 870a / 874a and bonding pad pairs 870b / 874b. In some embodiments, bonding pads 870 and / or 874 are flip chip bonding pads.
[0119] In one embodiment, the pattern of bonding pads 870 matches the pattern of bonding pads 874. In one embodiment, bonding pads 870 are bonded (e.g., flip chip bonded) to bonding pads 874. Thus, bonding pads 870, 874 electrically and physically couple memory die 302 to control die 304. In addition, bonding pads 870, 874 permit internal signal transmission between memory die 302 and control die 304. Thus, memory die 302 and control die 304 are bonded together. Although Fig. 8A One control die 304 is depicted bonded to one memory die 302 , but in one embodiment, one control die 304 is bonded to two memory dies 302 .
[0120] As used herein, "internal signaling" means signaling between the control die 304 and the memory die 302. Internal signaling permits circuitry on the control die 304 to control memory operations in the memory die 302. Thus, the bonding pads 870, 874 may be used for memory operation signaling. As used herein, "memory operation signaling" refers to any signal related to memory operations in the memory die 302. Memory operation signaling may include, but is not limited to, providing voltage, providing current, receiving voltage, receiving current, sensing voltage, and / or sensing current.
[0121] The bonding pads 870, 874 may be formed, for example, of copper, aluminum, and alloys thereof. A liner may be present between the bonding pads 870, 874 and the primary surface (882, 884). The liner may be formed, for example, of a titanium / titanium nitride stack. The bonding pads 870, 874 and the liner may be coated by vapor deposition and / or plating techniques. The bonding pads and liner together may have a thickness of 720 nm, but in other embodiments, this thickness may be greater or less.
[0122] Metal interconnects and / or vias may be used to electrically connect various elements in the die to bonding pads 870, 874. Several conductive paths that may be implemented via metal interconnects and / or vias are depicted. For example, sense amplifier 350 may be electrically connected to bonding pad 874b via path 812. There may be thousands of such sense amplifiers, paths, and bonding pads. It should be noted that BL does not necessarily form a direct connection to bonding pad 870b. Word line driver 860 may be electrically connected to bonding pad 874a via path 802. It should be noted that path 802 may include separate conductive paths for each word line driver 860(1) to 860(n). Likewise, there may be a separate bonding pad 874a for each word line driver 860(1) to 860(n). The word lines in block 2 of memory die 302 may be electrically connected to bonding pad 870a via path 804. In Fig. 8AIn FIG. 8 , there are “n” paths 804 for the corresponding “n” word lines in a block. There may be a separate pair of bond pads 870a, 874a for each path 804. Fig. 8A Further details of one embodiment of an integrated memory assembly 104 with metal interconnects and / or vias are depicted.
[0123] Figure 8B Another embodiment of an integrated memory assembly 104 is depicted in which a control die 304 can be used to control two memory dies 302a, 302b. The control die 304 has a plurality of bonding pads 874(a), 874(b) on a first major surface 884, such as in combination with Fig. 8A As discussed. Control die 304 has a plurality of bonding pads 876(a), 876(b) on second major surface 888. There may be "n" bonding pads 876(a) to deliver voltages from corresponding "n" word line drivers 860(1) to 860(n) to memory die 302b. Word line driver 860 may be electrically connected to bonding pad 876a via path 806. There may be one bonding pad 876b for each bit line associated with plane 830 on memory die 302b. Reference numeral 876 will be used to generally refer to bonding pads on major surface 888.
[0124] The second memory die 302b has several bonding pads 872a, 872b on the first main surface 886 of the second memory die 302b. There may be "n" bonding pads 872a to receive voltages from corresponding "n" word line drivers 860(1) to 860(n). The word lines in the plane 830 may be electrically connected to the bonding pads 872a via paths 808. There may be one bonding pad 872b for each bit line associated with the plane 830. Reference numeral 872 will be used to generally refer to the bonding pads on the main surface 886. It should be noted that there may be bonding pad pairs 872a / 876a and bonding pad pairs 872b / 876b. In some embodiments, bonding pads 872 and / or 876 are flip chip bonding pads.
[0125] In one embodiment, "n" word line drivers 860(1)-860(n) are shared between the two memory dies 302a, 302b. For example, a single word line driver may be used to provide voltage to a word line in the memory die 302a and a word line in the memory die 302b. However, the word line drivers 860 need not be shared between the memory dies 302a, 302b.
[0126] Fig. 9AFIG. 9 is a top view of a semiconductor wafer 935a from which a plurality of control dies 304 may be formed. The wafer 935a has many copies of the integrated circuits 903. In one embodiment, each of the integrated circuits 903 contains a control circuit 310 (see FIG. 3 ). Figure 3 ). In some embodiments, the wafer 935a is cut into semiconductor dies each containing one of the copies of the integrated circuit 903. Thus, many control semiconductor dies 304 may be formed from the wafer 935a. It should also be noted that even before the wafer 935a is cut, each region in which the integrated circuit 903 resides may be referred to as a control semiconductor die 304, as the term "control semiconductor die" is used herein.
[0127] Fig. 9B FIG. 9 is a top view of a semiconductor wafer 935b from which multiple memory dies 302 may be formed. Wafer 935b has many copies of integrated circuits 905. In one embodiment, each of the integrated circuits 905 contains a memory structure 326 (see FIG. Figure 3 ). In some embodiments, the wafer 935b is cut into semiconductor dies each containing one of the copies of the integrated circuit 905. Thus, many memory semiconductor dies 302 may be formed from the wafer 935b. It should also be noted that even before the wafer 935b is cut, each region in which the integrated circuit 905 resides may be referred to as a memory semiconductor die 302 as the term "memory semiconductor die" is used herein.
[0128] Semiconductor wafer 935 can be started as an ingot of monocrystalline silicon grown according to CZ, FZ or other processes. Semiconductor wafer 935 can be cut and polished on the main surface to provide a smooth surface. Integrated circuits 903, 905 can be formed on and / or in the main surface. Wafer 935 cutting into semiconductor dies can occur before or after bonding. In one embodiment, two wafers 935, 935b are bonded together. After the two wafers are bonded together, cutting is performed. Therefore, many integrated memory assemblies 104 can be formed by two wafers 935. In another embodiment, two wafers 935a, 935b are cut into semiconductor dies 304, 302. Subsequently, one of each of semiconductor dies 304, 302 is bonded together to form integrated memory assembly 104. Regardless of whether cutting occurs before or after bonding, it can be said that integrated memory assembly 104 contains control semiconductor die 304 and memory semiconductor die 302 bonded together.
[0129] The cutting of the wafer 935 into semiconductor dies may occur before or after bonding. In one embodiment, two wafers 935, 935b are bonded together. After the two wafers are bonded together, cutting is performed. Therefore, many integrated memory assemblies 104 may be formed by two wafers 935. In another embodiment, two wafers 935a, 935b are cut into semiconductor dies 304, 302. Subsequently, one of each of the semiconductor dies 304, 302 is bonded together to form an integrated memory assembly 104. Regardless of whether the cutting occurs before or after bonding, it can be said that the integrated memory assembly 104 contains a control semiconductor die 304 and a memory semiconductor die 302 bonded together.
[0130] As has been briefly discussed above, the control die 304 and the memory die 302 can be bonded together. Bond pads on each die 302, 304 can be used to bond the two dies together. Fig. 8A One example of a bond 870 on the memory die 302 and a bond 874 on the semiconductor die 304 is depicted. Fig.10 Depicts an example pattern of bonding pads on a planar surface of a semiconductor die. The semiconductor die may be memory die 302 or control die 304. The bonding pads may be either bonding pads 870 or 874, depending on the needs of the semiconductor die. Fig.10 As an example, 100,000 or more interconnects may be required between two of the semiconductor dies. To support this large number of electrical interconnects, the bonding pads may have a small area and spacing. In some embodiments, the bonding pads are flip chip bonding pads.
[0131] The semiconductor dies 302, 304 in the integrated memory assembly 104 can be bonded to each other by initially aligning the bonding pads 870, 874 on the respective dies 302, 304 with each other. Thereafter, the bonding pads can be bonded together by any of a variety of bonding techniques depending in part on the bonding pad size and bonding pad spacing (i.e., bonding pad pitch). The bonding pad size and pitch can in turn be dictated by the number of electrical interconnects required between the first semiconductor die 302 and the second semiconductor die 304.
[0132] In some embodiments, the bonding pads are directly bonded to each other in the so-called Cu to Cu bonding process without solder or other additional materials. In the Cu to Cu bonding process, the bonding pads are controlled to be highly flat and formed in a highly controlled environment, and the highly controlled environment is substantially free of environmental particles that may originally stay on the bonding pads and hinder tight bonding. Under such properly controlled conditions, the bonding pads are aligned and squeezed to form mutual bonding based on surface tension. Such bonding can be formed at room temperature, but heat can also be applied. In the embodiment using Cu to Cu bonding, the bonding pads can be about 5 μm square and spaced apart from each other at a spacing of 5 μm to 5 μm. Although this process is referred to as Cu to Cu bonding in this article, this term can also be applied even when the bonding pads are formed by materials other than Cu.
[0133] When the area of the bonding pad is small, it may be difficult to bond the semiconductor die together. The size of the bonding pad and the spacing between them can be further reduced by disposing a film layer on the surface of the semiconductor die containing the bonding pad. The film layer is disposed around the bonding pad. When the die are brought together, the bonding pads can be bonded to each other, and the film layers on the corresponding die can be bonded to each other. Such bonding techniques may be referred to as hybrid bonding. In an embodiment using hybrid bonding, the bonding pads may be approximately 5 μm square and spaced apart from each other at a spacing of 1 μm to 5 μm. Bonding techniques may be used to provide bonding pads with even smaller sizes and spacings.
[0134] Some embodiments may include a film on the surface of the die 302, 304. In the case where such a film is not initially provided, the space between the die may be underfilled with an epoxy or other resin or polymer. The underfill material may be applied as a liquid that subsequently hardens into a solid layer. This underfill step protects the electrical connections between the die 302, 304 and further holds the die together. Various materials may be used as the underfill material, but in an embodiment, the underfill material may be a Hysol epoxy resin from Henkel Corp., which has offices in California, USA.
[0135] In some embodiments, there is more than one control die 304 and more than one memory die 302 in the integrated memory assembly 104. In some embodiments, the integrated memory assembly 104 includes a stack of multiple control dies 304 and multiple memory dies 302. Fig.11AA side view of an embodiment of an integrated memory assembly 104 (e.g., a stack including control dies 304 and memory dies 302) stacked on a substrate 1102 is depicted. The integrated memory assembly 104 has three control dies 304a, 304b, 304c and three memory dies 302a, 302b, 302c. In some embodiments, there are far more than three memory dies 302 and far more than three control dies 304. In one embodiment, one of the control dies (e.g., 302a) is a master control die. In some embodiments, the master control die determines the memory die 302 schedule in order to implement a power / current plan for the integrated memory assembly 104.
[0136] Each control die 304 is attached (e.g., bonded) to at least one of the memory dies 302. Some of the bonding pads 870, 874 are depicted. There may be many more bonding pads. The space between the two dies 302, 304 bonded together is filled with a solid layer 1148, which may be formed of epoxy or other resins or polymers. This solid layer 1148 protects the electrical connections between the dies 302, 304 and further holds the dies together. Various materials may be used as the solid layer 1148, but in an embodiment, the solid layer 1148 may be a Hysol epoxy resin from Henkel Corporation, which has offices in California, USA.
[0137] The integrated memory assemblies 104 may be stacked, for example, in a stepped offset fashion, so that the bonding pads 1104 at each level are uncovered and accessible from above. Wire bonds 1106 connected to the bonding pads 1104 connect the control die 304 to the substrate 1102. The control die 304 may be provided over the width (i.e., from 1.25 to 1.5 m) of each control die 304. Fig.11A Several such wire bonds are formed.
[0138] Memory die through silicon vias (TSVs) 1112 may be used to route signals through the memory die 304. Control die through silicon vias (TSVs) 1114 may be used to route signals through the control die 302. In some embodiments, a master control die (e.g., 304a) communicates with slave control dies (e.g., 304b, 304c) via a communication link 608. In some embodiments, at least a portion of the communication link 608 extends through one or more memory die TSVs 1112 and one or more control die TSVs 1114.
[0139] The TSVs 1112, 1114 may be formed before, during, or after forming integrated circuits in the semiconductor dies 302, 304. The TSVs may be formed by etching holes through the wafer. The holes may then be lined with a barrier layer that prevents metal diffusion. The barrier layer may in turn be lined with a seed layer, and the seed layer may be plated with an electrical conductor such as copper, although other suitable materials may also be used, such as aluminum, tin, nickel, gold, doped polysilicon, and alloys or combinations thereof.
[0140] Solder balls 1108 may optionally be attached to contact pads 1110 on the lower surface of substrate 1102. Solder balls 1108 may be used to electrically and mechanically couple integrated memory assembly 104 to a host device, such as a printed circuit board. Solder balls 1108 may be omitted if integrated memory assembly 104 is to be used as an LGA package. Solder balls 1108 may form part of the interface between integrated memory assembly 104 and memory controller 102.
[0141] Fig. 11B A side view of an embodiment of an integrated memory assembly 104 stacked on a substrate 1102 is depicted. The integrated memory assembly 104 has three control dies 304d, 304e, 304f and three memory dies 302d, 302e, 302f. In some embodiments, there are far more than three memory dies 302 and far more than three control dies 304. In one embodiment, one of the control dies (e.g., 302d) is a master control die. In some embodiments, the master control die determines the memory die 302 schedule in order to implement a power / current plan for the integrated memory assembly 104. In this example, each control die 304 is bonded to at least one memory die 302. Optionally, the control die 304 may be bonded to two memory dies 302. For example, the control die 304e is bonded to both memory dies 302d and 304e. Likewise, the control die 304f is bonded to both memory dies 302e and 304f.
[0142] Some of the bonding pads 870, 874 are depicted. There may be many more bonding pads. The space between the two dies 302, 304 bonded together is filled with a solid layer 1148, which may be formed of epoxy or other resins or polymers. Fig.11A In the example, Fig. 11B The integrated memory assembly 104 in does not have a step offset. Memory die through silicon vias (TSVs) 1112 may be used to route signals through the memory die 302. Control die through silicon vias (TSVs) 1114 may be used to route signals through the control die 304.
[0143] Solder balls 1108 may optionally be attached to contact pads 1110 on the lower surface of substrate 1102. Solder balls 1108 may be used to electrically and mechanically couple integrated memory assembly 104 to a host device, such as a printed circuit board. Solder balls 1108 may be omitted if integrated memory assembly 104 is to be used as an LGA package.
[0144] Fig.12 is a perspective view of a portion of an example embodiment of a monolithic three-dimensional memory array that may include a memory structure 326 including a plurality of non-volatile memory cells. For example, Fig.12 A portion of a block of memory is shown. The depicted structure includes a set of bit lines BL positioned above a stack of alternating dielectric layers and conductive layers, wherein vertical material columns extend through the dielectric layers and the conductive layers. For exemplary purposes, one of the dielectric layers is labeled D and one of the conductive layers (also known as word line layers) is labeled W. The number of alternating dielectric layers and conductive layers may vary based on specific implementation requirements. A set of embodiments includes alternating dielectric layers and conductive layers between 108 and 304. An example embodiment includes 96 data word line layers, 8 selection layers, 6 dummy word line layers, and 110 dielectric layers. More or less than 108 to 304 layers may also be used. In one embodiment, the alternating dielectric layers and conductive layers are divided into four "fingers" or sub-blocks by local interconnects LI. FIG. 9 shows two fingers and two local interconnects LI. The source line layer SL is below the alternating dielectric layer and word line layer. Vertical columns of material (also referred to as memory holes) are formed in the stack of alternating dielectric and conductive layers. For example, one of the vertical columns / memory holes is labeled MH. It should be noted that in FIG. 9 , the dielectric layers are depicted as being see-through so that the reader can see the memory holes positioned in the stack of alternating dielectric and conductive layers. In one embodiment, a NAND string is formed by filling the vertical columns / memory holes with a material including a charge trapping material to form a vertical column of memory cells. Each memory cell can store one or more bits of data.
[0145] Fig.13A is a diagram of one embodiment of an integrated memory assembly 104. Fig.13A In one embodiment depicted in FIG. 3 , memory die 302 is bonded to control die 304. This bonding configuration is similar to Fig. 8A It should be noted that although gaps between adjacent pairs of dies are depicted, such gaps may be filled with epoxy or other resins or polymers. Fig.13A Additional details of one embodiment of path 352 are shown.
[0146] The memory die includes a memory structure 326. The memory structure 326 is adjacent to a substrate 1372 of the memory die 302. In some embodiments, the substrate 1372 is formed from a portion of a silicon wafer. In this example, the memory structure 326 includes a three-dimensional memory array. The memory structure 326 has Fig.12 There are several word line layers (WL) separated by dielectric layers. The dielectric layers are represented by the gaps between the word line layers. Thus, the word line layers and the dielectric layers form a stack. There may be more than Fig.13A As depicted in FIG. Fig.12 In an example of , there are several columns extending through the stack. One column 1302 is referred to in each stack by reference numeral 1302. The columns contain memory cells. For example, each column may contain a NAND string. There are several bit lines (BL) adjacent to the stack.
[0147] The word line driver 860 provides voltage in parallel to the word line 1342 in the memory die 302. The path from the word line driver 860 to the word line 1342 includes a conductive path 1332, a bonding pad 874a, a bonding pad 870a, and a conductive path 1334. In some embodiments, the conductive paths 1332, 1334 are referred to as a path pair. The conductive paths 1332, 1334 may each include one or more through-holes (which may extend vertically relative to the main surface of the die) and one or more metal interconnects (which may extend horizontally relative to the main surface of the die). The conductive paths 1332, 1334 may include transistors or other circuit elements. In one embodiment, a transistor can be used to effectively cut off or close a path. Other word line drivers ( Fig.13A The bonding pads 874a and 870a are provided with voltages to other word lines (not depicted in FIG. 1 ). Therefore, there are additional bonding pads 874a and 870a in addition to the bonding pads 874a and 870a. As is known in the art, the bonding pads may be formed, for example, of copper, aluminum, and alloys thereof.
[0148] Sense amplifier 350 communicates with the bit line in memory die 302. The path from sense amplifier 350 to the bit line includes conductive path 1352, bonding pad 874b, bonding pad 870b, and conductive path 1354. In some embodiments, conductive paths 1352, 1354 are referred to as a path pair. Conductive paths 1352, 1354 may include one or more through-holes (which may extend vertically relative to the main surface of the die) and one or more metal interconnects (which may extend horizontally relative to the main surface of the die). The metal interconnects may be formed of various conductive metals including, for example, copper and copper alloys as known in the art, and the through-holes may be lined with and / or filled with various conductive metals including, for example, tungsten, copper, and copper alloys as known in the art. Conductive paths 1352, 1354 may include transistors or other circuit elements. In one embodiment, transistors may be used to effectively cut off or close paths.
[0149] The control die 304 has a substrate 1376 which may be formed from a silicon wafer. The sense amplifier 350, the word line driver 860, and other circuits 1320 may be formed on and / or in the substrate 1376. The circuits 1320 may include the control circuit 310 (see Figure 3 ) Some or all of. In some embodiments, the sense amplifier 350, the word line driver 860 and / or other circuits 1320 include CMOS circuits.
[0150] There are external signal paths that allow circuits on the control die 304 to communicate with entities external to the integrated memory assembly 104, such as the memory controller 102. Thus, the circuits 1320 on the control die 304 can communicate with, for example, the memory controller 102 (see Figure 3 ) communication. Optionally, circuits on the control die 304 can communicate with, for example, the host 120. The external path includes a through hole 1358 in the control die 304, a bonding pad 874c, a bonding pad 870c, a through silicon via (TSV) 1360, and an external pad 1378. The TSV 1360 extends through the substrate 1372.
[0151] The TSVs 1360 may be formed before, during, or after forming integrated circuits in the semiconductor dies 302, 304. The TSVs may be formed by etching holes through the wafer. For example, the holes may be etched through the substrate 1372. The holes may also be etched through material adjacent to the wafer. The holes may then be lined with a barrier layer that prevents metal diffusion. The barrier layer may in turn be lined with a seed layer, and the seed layer may be plated with an electrical conductor such as copper, although other suitable materials may also be used, such as aluminum, tin, nickel, gold, doped polysilicon, and alloys or combinations thereof.
[0152] right Fig.13AMany modifications to the embodiment depicted in are possible. One modification is to locate the sense amplifier 350 on the memory die 302.
[0153] Fig. 13B is a diagram of one embodiment of an integrated memory assembly 104. This bonding configuration is similar to Figure 8B An embodiment described in . Fig. 13B The configuration in Fig.13A Therefore, similar reference numerals are used for Fig.13A The memory die 302 in Fig. 13B The memory die 302a in Fig. 13B In the embodiment depicted in , a first memory die 302a is bonded to a control die 304, and the control die 304 is bonded to a second memory die 302b. It should be noted that while gaps between adjacent pairs of dies are depicted, such gaps may be filled with epoxy or other resins or polymers.
[0154] Each memory die 302a, 302b includes a memory structure 326. Memory structure 326a is adjacent to substrate 1372 of memory die 302a. Memory structure 326b is adjacent to substrate 1374 of memory die 302b. In some embodiments, substrates 1372, 1374 are formed from a portion of a silicon wafer. In this example, memory structures 326 each include a three-dimensional memory array.
[0155] The word line driver 860 provides voltage in parallel to the first word line 1342 in the memory die 302a and the second word line 1344 in the memory die 302b. The path from the word line driver 860 to the second word line 1344 includes the conductive path 1332, the through silicon via (TSV) 1368, the bonding pad 876a, the bonding pad 872a and the conductive path 1336. Fig. 13B (not depicted) provides voltage to other word lines.
[0156] Sense amplifier 350a communicates with a bit line in memory die 302a. The path from sense amplifier 350a to the bit line includes conductive path 1352, bonding pad 874b, bonding pad 870b, and conductive path 1354. Sense amplifier 350b communicates with a bit line in memory die 302b. The path from sense amplifier 350b to the bit line includes conductive path 1354, TSV 1356, bonding pad 876b, bonding pad 872b, and conductive path 1348.
[0157] right Fig. 13BMany modifications of the embodiment depicted in are possible. One modification is to locate sense amplifier 350a on the first memory die 302a and sense amplifier 350b on the second memory die 302b.
[0158] Fig.14 is a flow chart describing one embodiment of a process 1400 for programming a NAND string of memory cells organized into an array. In one example embodiment, the control circuit 310 discussed above is used to perform a program on the integrated memory assembly 104. Fig.14 For example, the process may be performed at the direction of the processor 312 Fig.14 The process includes a plurality of loops, each of which includes a programming phase (eg, steps 1404 to 1408) and a verification phase (eg, steps 1410 to 1418).
[0159] In many implementations, the magnitude of the programming pulses increases by a predetermined step size with each successive pulse. Fig.14 In step 1402, the programming voltage (Vpgm) is initialized to a starting magnitude (eg, ˜12 to 16 V or another suitable level) and a program counter PC maintained by processor 312 is initialized at 1.
[0160] In one embodiment, a group of memory cells selected to be programmed (referred to herein as selected memory cells) are programmed in parallel and are all connected to the same word line (selected word line). Other memory cells not selected for programming (unselected memory cells) may also be connected to the selected word line. That is, the selected word line will also be connected to the memory cells that should be prohibited from programming. In addition, as the memory cells reach their intended target data states, further programming of the memory cells will be prohibited. Those NAND strings (e.g., unselected NAND strings) that contain memory cells connected to the selected word line that should be prohibited from programming have their channels boosted to prohibit programming. When the channel has a boosted voltage, the voltage difference between the channel and the word line is not large enough to cause programming. To assist in boosting, in step 1404, the memory system will precharge the channels of the NAND strings that contain memory cells connected to the selected word line that should be prohibited from programming.
[0161] In step 1406, NAND strings containing memory cells connected to the selected word line that should be inhibited from programming have their channels boosted to inhibit programming. Such NAND strings are referred to herein as "unselected NAND strings". In one embodiment, the unselected word lines receive one or more boost voltages (e.g., ~7 to 11 volts) to perform the boosting scheme. The program inhibit voltage is applied to the bit lines coupling the unselected NAND strings.
[0162] In step 1408, a programming pulse Vpgm of a programming signal is applied to the selected word line (the word line selected for programming). In one embodiment, if a memory cell on a NAND string should be programmed, the corresponding bit line is biased at a program enable voltage. Such a NAND string is referred to herein as a "selected NAND string".
[0163] In step 1408, programming pulses are applied in parallel to all memory cells connected to the selected word line so that all memory cells connected to the selected word line are programmed in parallel (unless the memory cells are inhibited from programming). That is, the memory cells are programmed at the same time or during overlapping times (both are considered parallel). In this way, all memory cells connected to the selected word line will have their threshold voltages changed in parallel unless programming is inhibited.
[0164] In step 1410, memory cells that have reached their target state are locked out from further programming. Step 1410 may include verifying at one or more verify reference levels. In one embodiment, the verify process is performed by testing whether the threshold voltage of the memory cells selected for programming has reached the appropriate verify reference voltage.
[0165] In step 1410, the memory cell may be locked out after the memory cell has been verified (passed a test of Vt) that the memory cell has reached its target state.
[0166] If, in step 1412, it is determined that all memory cells have reached their target threshold voltage (PASS), then the programming process is complete and successful because all selected memory cells were programmed and verified to their target states. A "PASS" status is reported in step 1414. Alternatively, if, in 1412, it is determined that not all memory cells have reached their target threshold voltage (FAIL), then the programming process continues to step 1416.
[0167] In step 1416, the memory system counts the number of memory cells that have not yet reached their corresponding target threshold voltage distribution. That is, the system counts the number of memory cells that have failed to reach their target state so far. This counting can be performed by the processor 312, the memory controller 102, or other logic. In one embodiment, each of the sense blocks will store the state (pass / fail) of its corresponding cell. In one embodiment, there is a total count that reflects the total number of memory cells currently being programmed that failed the last verification step. In another embodiment, a separate count is saved for each data state.
[0168] In step 1418, it is determined whether the count from step 1416 is less than or equal to a predetermined limit. In one embodiment, the predetermined limit is the number of bits that can be corrected by an error correction code (ECC) during the read process of the memory cell page. If the number of cells that did not pass is less than or equal to the predetermined limit, the programming process can be stopped and a "pass" status is reported in step 1414. In this case, enough memory cells are correctly programmed so that the few remaining memory cells that have not been fully programmed can be corrected using ECC during the read process. In some embodiments, the predetermined limit used in step 1418 is less than the number of bits that can be corrected by an error correction code (ECC) during the read process, thereby allowing for future / additional errors. When programming less than all memory cells of a page, or comparing counts of only one data state (or less than all states), then the predetermined limit can be a portion (proportional or non-proportional) of the number of bits that can be corrected by ECC during the read process of the memory cell page. In some embodiments, the limit is not predetermined. Alternatively, the limit is changed based on the number of errors counted for the page, the number of program-erase cycles performed, or other criteria.
[0169] If the number of failed memory cells is not less than a predetermined limit, the programming process continues at step 1420 and the program counter PC is checked against a program limit (PL). Examples of program limits include 6, 12, 16, 19, and 30; however, other values may be used. If the program counter PC is not less than the program limit PL, the programming process is deemed to have failed and a failed status is reported in step 1424. If the program counter PC is less than the program limit PL, the process continues at step 1422, during which the program counter PC is incremented by 1 and the programming voltage Vpgm is increased to the next magnitude. For example, the next pulse will have a magnitude that is one step larger (e.g., a step size of 0.1 to 1.0 volts) than the previous pulse. After step 1422, the process loops back to step 1404 and another programming pulse is applied to the selected word line so that the programming process is performed. Fig.14 Another iteration of the programming process (steps 1404 to 1422).
[0170] In some embodiments, the control die 304 determines how much power or current will be drawn through its memory die 302 in the upcoming time period based on how much power / current will be drawn in various steps in the process 1400. The time period may cover multiple steps of the process 1400. In some embodiments, the main control die 304a selects which memory die 304 to perform memory operations during the time period. In some embodiments, this may cause the control die 302 to temporarily pause the process 1400. For example, the process 1400 may be temporarily paused between steps 1408 and 1410 (i.e., between the programming phase and the verification phase). The process 1400 may be temporarily paused between the verification phase and the next programming phase (e.g., before step 1404). In some embodiments, some of the steps in the process 1400 are considered atomic operations, which should not be stopped after they are in progress. In one embodiment, steps 1404 to 1406 are performed as atomic operations. That is, after starting step 1404, the operation is not paused until step 1408 is completed. In some embodiments, being able to stop operations during process 1400 allows for fine grain control of power / current used in integrated memory assembly 104. Such fine grain control may not be possible or at least not feasible if memory controller 102 is to regulate power usage.
[0171] At the end of a successful programming process, the threshold voltage of the memory cell should be within one or more threshold voltage distributions for programmed memory cells or within a threshold voltage distribution for erased memory cells, as appropriate. Fig.15A An example threshold voltage distribution of a memory array is shown when each memory cell stores three data bits. However, other embodiments may use other data capacities per memory cell (e.g., one, two, four, or five data bits per memory cell). FIG. 15A shows eight threshold voltage distributions corresponding to eight data states. The first threshold voltage distribution (data state) Er represents an erased memory cell. The other seven threshold voltage distributions (data states) A to G represent programmed memory cells and are therefore also referred to as programmed states. Each threshold voltage distribution (data state) corresponds to a predetermined value of a set of data bits. The specific relationship between the data programmed into the memory cell and the threshold voltage level of the cell depends on the data encoding scheme adopted by the cell. In one embodiment, a Gray code is used to assign data values to a threshold voltage range so that if the threshold voltage of the memory is erroneously shifted to its adjacent physical state, only one bit will be affected.
[0172] Fig.15ASeven read reference voltages, VrA, VrB, VrC, VrD, VrE, VrF, and VrG, are shown for reading data from memory cells. By testing (e.g., performing a sensing operation) whether the threshold voltage of a given memory cell is above or below the seven read reference voltages, the system can determine which data state (i.e., A, B, C, D, ...) the memory cell is in.
[0173] Fig.15A Also shown are seven verification reference voltages VvA, VvB, VvC, VvD, VvE, VvF, and VvG. In some embodiments, when programming memory cells to data state A, the system will test whether those memory cells have a threshold voltage greater than or equal to VvA. When programming memory cells to data state B, the system will test whether the memory cells have a threshold voltage greater than or equal to VvB. When programming memory cells to data state C, the system will determine whether the memory cells have their threshold voltage greater than or equal to VvC. When programming memory cells to data state D, the system will test whether those memory cells have a threshold voltage greater than or equal to VvD. When programming memory cells to data state E, the system will test whether those memory cells have a threshold voltage greater than or equal to VvE. When programming memory cells to data state F, the system will test whether those memory cells have a threshold voltage greater than or equal to VvF. When programming memory cells to data state G, the system will test whether those memory cells have a threshold voltage greater than or equal to VvG. Fig.15A Also shown is Vev, which is the voltage level used to test whether the memory cell has been properly erased.
[0174] In one embodiment, referred to as full sequence programming, memory cells may be programmed directly from erased data state Er to any of programmed data states A to G. For example, a group of memory cells to be programmed may first be erased so that all memory cells in the group are in erased data state Er. Subsequently, a programming process is used to program memory cells directly to data states A, B, C, D, E, F, and / or G. For example, while some memory cells are programmed from data state ER to data state A, other memory cells are programmed from data state Er to data state B and / or from data state Er to data state C, and so on. Fig.15A The arrows represent full sequence programming. In some embodiments, data states A through G may overlap, where the control die 304 and / or memory controller 102 rely on error correction to identify the correct data being stored.
[0175] The techniques described herein may also be used with other types of programming other than full sequence programming, including but not limited to multi-level / multi-phase programming. In one embodiment of multi-level / phase programming, all memory cells ending with any of data states D to G are programmed to an intermediate state that is no higher than D in the first phase. Memory cells ending with any of data states Er to C do not receive programming in the first phase. In the second phase, memory cells ending with data states B or C are programmed to a state no higher than B; memory cells ending with data states F or G are programmed to a state no higher than F. In the third phase, the memory cells are programmed to their final states. In one embodiment, the first page is programmed in the first phase, the second page is programmed in the second phase, and the third page is programmed in the third phase. Herein, after the pages have been programmed into groups of memory cells, the memory cells can be read back to retrieve the pages. Therefore, the intermediate states associated with multi-phase programming are considered programmed states herein.
[0176] Generally speaking, during verify operations and read operations, a selected word line is connected to a voltage (an example of a reference signal) whose level is specified for each read operation (see, for example, Fig.15A The read compare levels VrA, VrB, VrC, VrD, VrE, VrF, and VrG) or the levels of the voltages specified for each verification operation (see, for example, Fig.15A Verification target levels VvA, VvB, VvC, VvD, VvE, VvF, and VvG) are measured to determine whether the threshold voltage of the associated memory cell has reached such a level. After the word line voltage is applied, the conduction current of the memory cell is measured to determine whether the memory cell is turned on (conducting current) in response to the voltage applied to the word line. If the conduction current is measured to be greater than a certain value, it is assumed that the memory cell is turned on and the voltage applied to the word line is greater than the threshold voltage of the memory cell. If the conduction current is not measured to be greater than the certain value, it is assumed that the memory cell is not turned on and the voltage applied to the word line is not greater than the threshold voltage of the memory cell. During the read or verify process, the unselected memory cells are provided with one or more read pass voltages (also called bypass voltages) at their control gates so that these memory cells will operate as conduction gates (e.g., conduct current regardless of whether these memory cells are programmed or erased).
[0177] There are many ways to measure the conduction current of a memory cell during a read or verify operation. In one example, the conduction current of a memory cell is measured by the rate at which the conduction current discharges or charges a dedicated capacitor in a sense amplifier. In another example, the conduction current of a selected memory cell allows (or does not allow) the NAND string containing the memory cell to discharge the corresponding bit line. The voltage on the bit line is measured after a period of time to see if the bit line has discharged. It should be noted that the techniques described herein can be used with different methods known in the art for verification / reading. Other read and verify techniques known in the art can also be used.
[0178] Fig. 15B Depicted are threshold voltage distributions and a page mapping scheme when each memory cell stores four bits of data. Fig. 15B It is depicted that there may be some overlap between data states S0 and S15. Overlap may occur due to factors such as memory cells losing charge (and therefore threshold voltage drops). Program disturb may unintentionally increase the threshold voltage of a memory cell. Likewise, read disturb may unintentionally increase the threshold voltage of a memory cell. Over time, the position of the threshold voltage distribution may change. Such changes may increase the bit error rate, thereby extending the decoding time or even making decoding impossible.
[0179] As mentioned, Fig. 15B An example is depicted where four bits are stored per memory cell. Thus, four pages can be stored in a set of memory cells. Fifteen read reference levels (Vrl through Vrl5) are depicted. A set of memory cells can be connected to the same word line.
[0180] Fig.16 and 17 Further details of one embodiment of the programming phase of a programming operation are described. Fig.16 Describe the application can be used Fig.14 Flowchart of one embodiment of a programming pulse and programming voltage for step 1408. In one embodiment, this process is used to program 3D NAND. Fig.17 Shown in Fig.16 In one embodiment, the processor 312 controls the timing of various signals during the process of Fig.17 The timing and magnitude of the signals in .
[0181] At step 1622, a low voltage is applied to the source line. Fig.17, the common source line SL is raised to a low voltage Vsrc (e.g., about 1.0V to 1.5V) at time t0, where the low voltage state remains unchanged for the duration of the programming phase. In some embodiments, the source line SL remains at 0V. The source select line SGS remains at 0V to place the source select transistor in isolation mode. As an example, this can be about 1.0V to 1.5V. It should be noted that Fig.17 The other signals in are at 0 V at time t0.
[0182] At step 1624, data is applied to the bit lines. In one embodiment, the processor 312 controls the sense amplifier 350 to apply the data to the bit lines. The data may be applied to the bit lines based on whether or not a memory cell in the NAND string that is in the current programming phase is being programmed. If the memory cell is being programmed, its associated bit line has a program enable voltage (e.g., Vbl_Program_Enable) applied to it. If the memory cell is not being programmed, its associated bit line has a program inhibit voltage (e.g., Vbl_Inhibit) applied to it.
[0183] See also Fig.17 , line 1752 shows the bit line voltage at 0V for a NAND string (in the set that is part of the current programming phase) having a memory cell at a selected word line WLn to be programmed. Line 1750 shows the bit line voltage at V for a NAND string (in the set that is part of the current programming phase) having a cell at WLn that is inhibited from programming. DD The bit line voltage under V DD = represents program inhibit and is often referred to as the program inhibit voltage. Note that line 1752 represents one instance of V_Program_Enable and line 1750 represents one instance of Vbl_Inhibit.
[0184] Step 1626 is to apply Vsgds to the selected drain side select line. Step 1628 is to apply Vsgdu to the unselected drain side select line. It should be noted that Fig.17 Two SGD lines are shown. Line SGD refers to the drain side select line that has at least one NAND string currently being programmed. Line SGDu refers to the drain side select line that has no NAND string currently being programmed.
[0185] See also Fig.17 At time t3, the drain side select line SGD is raised to Vsgds, and the drain side select line SGDu is raised to Vsgdu. The voltage Vsgds turns on the drain side select transistors for the NAND string with the memory cell being programmed. It should be noted that at time t3, Vbl_Program_Enable is being applied to the selected bit line BL0.
[0186] Note that at this time, Vsgds may also be applied to the control gates of the drain side select transistors of the NAND strings in which no memory cells should be programmed. Note, however, that at time t3, Vbl_Inhibit is applied to the unselected bit line BL1.
[0187] See again Fig.17 At time t3, the drain side select line SGDu is set to Vsgdu. The voltage Vsgdu should keep the associated drain side select transistor off.
[0188] At step 1636, the boost voltage (eg, V PASS ) is applied to the unselected word lines. It should be noted that V PASS The magnitude of does not need to be the same for each of the unselected word lines. There are various boosting schemes that can be used. Examples of boosting schemes include, but are not limited to, self-boosting (SB), local self-boosting (LSB), and erased area self-boosting (EASB).
[0189] See also Fig.17 At time t4, boosting / programming begins. At time t4, a boost voltage is applied to various word lines. In one embodiment, each unselected word line receives a boost voltage V PASS In one embodiment, the voltage applied to the unselected word lines depends on their position relative to the selected word line.
[0190] In one embodiment, a pass voltage is applied to at least a subset of unselected word lines during a programming phase of a programming operation. The pass voltage is typically less than the programming voltage. As an example, the pass voltage may be 10 volts. However, the pass voltage may be higher or lower. The pass voltage may assist in boosting the channel of the memory cell. In one embodiment, capacitive coupling between the control gate and the channel boosts the channel potential. Thus, the pass voltage applied to the control gate may be used to boost the channel potential.
[0191] At step 1638, the programming voltage V PGM Then applied to the selected word line.By applying the boost voltage and boosting the inhibited NAND string channel, program disturb of unselected memory cells on the selected word line is prevented or reduced.
[0192] See also Fig.17 , at time t5, the programming voltage V PGM Applied to the selected word line WLn. Because the disabled NAND string channel is boosted when the programming voltage is applied, the unselected memory cells at WLn for the disabled NAND string will not be programmed. The boosted channel region voltage reduces the potential on the tunneling dielectric region of those memory cells, thereby preventing any unintentional programming.
[0193] At step 1640, the word lines, bit lines, source lines, and drain select lines are lowered to 0V, thereby marking the completion of one programming iteration (eg, one programming phase). Fig.16 The steps can be applied to Fig.14 Each iteration of the process of increasing the programming voltage with each iteration. However, it is not required that the programming voltage be increased in magnitude with each iteration.
[0194] See also Fig.17 , at time t6, the word line is lowered to 0 V. The source and bit lines are then lowered to 0 V at time t7. The drain select line SGD is lowered to 0 V at time t8.
[0195] Fig.18 A flow chart describing an embodiment of a verify operation of a programming operation performed to verify whether a memory cell has been programmed to its target data state is depicted. The process may be performed under the control of processor 312. Process 1800 may be used during step 1410 of process 1400.
[0196] In step 1802, a pass voltage is applied to the unselected word lines to make the unselected memory cells on the NAND string conductive, which enables the system to test whether the selected memory cells are conductive in response to a verify reference voltage. This pass voltage is generally referred to as Vread. In step 1804, an appropriate verify reference voltage (also referred to as Vcgr) is applied to the selected word line (e.g., VvA, VvB, etc.).
[0197] Step 1806 includes sensing the selected memory cell. In one embodiment, the trip latch 468 in the sense amplifier 350 connected to the bit line for one of the selected memory cells may trip in response to the reference voltage applied to the selected word line. The corresponding output will be provided from the sense amplifier 350 to the latch manager 482 via the data bus 454. In one embodiment, when a memory cell trips, the processor 312 is notified.
[0198] In some embodiments, the sensing in step 1806 includes the following. In one example embodiment, a capacitor in a sense amplifier is charged. The bit line is connected to the capacitor to allow the bit line to discharge the capacitor in response to the reference voltage in step 1804. After a predetermined period of time called the "integration time" or "gating time", the voltage of the capacitor is sampled to see if the corresponding memory cell is conductive. If the memory cell is conductive in response to Vcgr, then the trip latch 468 is tripped, indicating that the threshold voltage of the memory cell is less than Vcgr. There are many other techniques for sensing memory cells.
[0199] Step 1808 is to determine if there is another verify reference level applied to the selected word line. Likewise, a determination will be made whether the memory cell trips in response to the reference voltage applied to the selected word line.
[0200] Based on the reference level (if present), the control die 304 will determine if the memory cell has reached its target state when the trip latch 468 in the sense amplifier 350 trips. For example, if the memory cell is being programmed to the C data state, the control die 304 determines if the threshold voltage of the memory cell is at least VvC.
[0201] The memory cells that passed verify are locked out from further programming in step 1810. In some embodiments, the voltage applied to the bit lines during the programming phase is used to control whether the memory cells are locked out.
[0202] An embodiment of the memory system 100 has a main control die 304a that regulates the power / current usage of the integrated memory assembly 104 . Fig.19 A flow chart depicting one embodiment of a process 1900 for regulating power / current usage of an integrated memory assembly 104. The process 1900 occurs in the integrated memory assembly 104.
[0203] Step 1902 includes the master control die 304a analyzing the power / current usage profile of each memory die 302 in the integrated memory assembly 104 in an upcoming time period. The power / current usage profile is an estimate of the power / current usage of memory operations expected in the near future. In one embodiment, each control die 304 determines the power / current usage profiles of the one or more memory die 302 it controls and sends those profiles to the master control die 304a. Reference will now be made to Fig. 20 , to discuss an example current usage profile of the memory die 302. In some embodiments, each slave control die 304b is based on Fig. 20 The usage profile is sent to the main control die 304a based on the example current usage profile depicted in FIG.
[0204] Fig. 20 Depicted are several graphs showing examples of current (e.g., Icc) usage profiles for several memory dies 302. Graph 2002 depicts the current usage profile for a memory die referred to as Die0. Graph 2004 depicts the current usage profile for a memory die referred to as Die1. Graph 2006 depicts the current usage profile for a memory die referred to as Dien.
[0205] For illustration purposes, each graph 2002, 2004, 2006 covers three time periods (Time Period 1, Time Period 2, Time Period 3). Each time period is for the future. Thus, the current usage profile is an estimate or prediction of current usage in each upcoming time period. In one embodiment, the estimated current usage is an estimate of the peak current consumption during the time period. In one embodiment, the estimated current usage is an estimate of the average current consumption during the time period.
[0206] exist Fig. 20 In the example of , the estimated current usage of die 0 is represented by line 2008a in time period 1, by line 2008b in time period 2, and by line 2008c in time period 3. The estimated current usage of die 1 is represented by line 2010a in time period 1, by line 2010b in time period 2, and by line 2010c in time period 3. The estimated current usage of die n is represented by line 2012a in time period 1, by line 2012b in time period 2, and by line 2012c in time period 3. Depicting the estimated current usage as a constant value in a particular time period simplifies the analysis of the estimated current usage. Therefore, the control die 304 can simply report one value to the main control die in each time period. However, the estimated current usage can have more than one value per time period. For example, the estimated current usage can include both the peak and average current usage in each time period.
[0207] There may be more or less than three time periods in the power usage profile of each memory die 302. The length of the time period may be set relatively small to help provide fine-grained control. For example, the time period may be about the time it takes to perform the program verification phase of a programming operation.
[0208] Reference again Fig.19 , step 1904 includes determining a plan that satisfies the power criteria for the integrated memory assembly 104 based on the analysis of step 1902. In one embodiment, the master control die 304a determines a power plan for executing memory operations in the memory die 302 to keep the combined power usage of the integrated memory assembly 104 within the power budget. In one embodiment, the power plan includes a die scheduling plan for each memory die 302. In one embodiment, the power criteria include the peak current consumption of the integrated memory assembly 104. As described above, adjusting the peak current can cause the adjustment of the peak power. In one embodiment, the master control die 304a attempts to maximize the current usage in each of the time periods while maintaining within the maximum allowed current for the integrated memory assembly 104. Reference will now be made to Fig.21To illustrate staying within the maximum allowed current for the integrated memory assembly 104.
[0209] Fig.21 A graph is depicted to illustrate one example of combined current usage of all memory dies 302 in an integrated memory assembly 104. Dashed line 2102 represents the maximum allowed current for the integrated memory assembly 104. In one embodiment, dashed line 2102 represents the maximum allowed peak current for the integrated memory assembly 104. Line 2104a represents the total current consumption for all memory dies 302 in the integrated memory assembly 104 during time period 1. Line 2104b represents the total current consumption for all memory dies 302 in the integrated memory assembly 104 during time period 2. Line 2104c represents the total current consumption for all memory dies 302 in the integrated memory assembly 104 during time period 3. It should be noted that lines 2104a, 2104b, and 2104c are based on current profiles (see Fig. 20 ). In one embodiment, these are the peak current draws for the integrated memory assembly 104 in each time period.
[0210] In some embodiments, the master control die 304a selects which of the memory semiconductor dies are performing memory operations or phases of memory operations in time periods to maximize the combined power usage of the integrated memory assembly 104 while keeping the combined power / current usage of the integrated memory assembly 104 within the power / current budget. In some embodiments, the master control die 304a determines the combination of memory semiconductor dies 302 performing phases of programming operations in time periods in order to maximize the power / current usage of the integrated memory assembly 104 in each time period while keeping within the power / current usage limit for the integrated memory assembly 104.
[0211] Also refer to Fig.19 1, step 1906 includes the master control die 304a sending die schedules for multiple time periods to the slave control die 304b. Each die schedule indicates when a memory operation or a phase of a memory operation is to be performed. In one embodiment, the common clock signal (see Figure 6 ) is used to coordinate operations or synchronize operations. For example, the die schedule may indicate which clock cycle the programming phase should start, which clock cycle the programming verification phase should start, etc. It should be noted that the master control die 304a also controls the memory die 302. Therefore, the master control die 304a can generate its own die schedule. In one embodiment, the master control die 304a sends the die schedule by means of the communication link 608, which avoids the use of the memory interface 322.
[0212] Step 1908 includes operating the memory die 302 in each time period according to the die schedule. In some embodiments, each control semiconductor die 304 determines when to perform each memory operation or phase of a memory operation based on a common clock signal. It should be noted that in step 1908, each control semiconductor die communicates with the memory structure 326 in the memory semiconductor die 302 through bonding pads (e.g., 870, 872, 874, 876) to perform a memory operation (or phase of a memory operation) at a specified time period.
[0213] As described above, the master control die 304a determines which memory dies 302 will operate in the upcoming time period. It is possible that some memory dies 302 will become starved. Therefore, in some embodiments, the priority of the memory dies 302 is increased to prevent starvation. Fig. 22 A flow chart depicting one embodiment of a process for the master control die 304a to develop a die scheduling plan. The process may be used to Fig.19 Step 1904.
[0214] Step 2202 includes selecting a memory die 302 to operate in the next time period based on the usage profile and the priority. An example of a usage profile is Fig. 20 22. Initially, each memory die 302 is assigned the same default priority. Step 2202 results in selecting a set of memory die 302 to operate in a time period such that the power / current criteria for that time period are met.
[0215] Step 2204 includes raising the priority of the memory die 302 that was not selected in step 2202. Step 2206 includes determining whether the master control die 304a is to be considered for more time periods for planning. If so, control passes to step 2202. In step 2202, the master control die 304a again selects a memory die 302 to operate in the next time period based on the usage profile and priority. Since the priority of the previously unselected memory die is raised, such memory die 302 has a greater chance of selection in step 2202. After the memory die 302 has been selected in step 2202, its priority can be reset to the default priority. After all time periods have been considered, the process ends.
[0216] As described above, in some embodiments, the control die 304 estimates the power / current usage profile for the upcoming time period. In some embodiments, these estimates are based on information stored by each control die 304 indicating how much power or current is consumed by various memory operations or their phases. Fig.23 in order to explain how such information may be determined. Fig.23 Two graphs showing the current usage of one programming loop are depicted. Graph 2302 depicts the voltage on the selected word line versus time. Graph 2304 depicts the current within the programming loop versus time. This current refers to the current drawn by the memory die 302. The current usage is a result of many other factors besides the voltage applied to the selected word line. During the programming phase, the programming voltage (Vpgm) is applied to the selected word line. The voltage Vpass is the boost voltage applied to all word lines during programming (see Fig.17 During the verify phase, one or more verify reference voltages are applied to the selected word line. Fig.23 In the figure, VrA, VrB and VrC are depicted. These can be Fig.15A . It should be noted that all verification reference voltages need not be applied in each programming loop. For example, in an early programming loop, higher reference voltages (e.g., VvD, VvE, VvF, and VvG) may be skipped.
[0217] There may be some variation in current usage from one memory die 302 to the next memory die. In some embodiments, the current graph 2304 is determined on a per memory die 302 basis. This may be determined offline or during use. For example, a current graph 2304 for each memory die 302 may be determined. The control die 304 may access data such as peak (or average) current usage for various phases of memory operation. These data may be stored in a non-volatile memory such as in the memory die 302. For example, the stored data may indicate peak current usage in the programming phase, peak current usage in the verification phase, peak current usage in a read operation, etc. Due to factors such as the number of verification reference voltages applied to the selected word line (and therefore the number of states sensed), there may be some variation in current usage from one programming loop to the next programming loop. Other factors may affect the current usage from one programming loop to the next programming loop. Therefore, the control die 302 may store current usage data for each loop of the programming operation.
[0218] In some embodiments, current usage is the primary concern during programming. However, current usage for other operations, such as erase and read, may also be determined. In some embodiments, an erase operation includes an erase phase followed by an erase verify phase. Thus, an erase operation may include multiple phases similar to a programming operation.
[0219] Fig.24A flow chart depicting one embodiment of a process 2400 for a control die 304 that determines a current usage profile. The current usage profile may be sent to a master control die 304a to allow the master control die to determine a power / current plan. Step 2402 includes the control die 302 determining a memory operation (or phase thereof) to be performed in a time period. In process 2400, it is assumed that the time period is small enough so that no more than one memory operation is performed per time period. However, the time period may be of any size such that, in some embodiments, more than one memory operation may be performed in a time period. In some cases, a memory operation will take multiple time periods to complete.
[0220] Step 2404 includes determining the peak current usage in a time period given the memory operation (or phase thereof) to be performed. Fig.23 , if the programming phase of a programming operation is to be performed, then the peak current for the programming phase can be based on graph 2304. Likewise, if the verify phase of a programming operation is to be performed, then the peak current for the verify phase can be based on graph 2304. As described above, control die 302 can store a table or the like containing data indicating how much current to use for different phases of a programming operation.
[0221] Step 2406 includes determining whether another time period is to be analyzed. If so, steps 2402 to 2404 are performed again for the memory operation (or phase thereof) to be performed in the next time period. After a sufficient number of time periods have been analyzed, in step 2408, the control die 302 provides the current usage profile to the main control die 304a. In one embodiment, the current usage profile includes an estimate of the peak current for each phase of the programming operation to be performed.
[0222] Fig.25 A flow chart depicting one embodiment of a process 2500 for a slave control die 304b performing memory operations according to a die scheduling plan provided by a master control die 304a. The process is described with respect to a programming operation, but the process may be modified for other memory operations. In step 2502, the slave control die 304b receives the die scheduling plan from the master control die 304a. In some embodiments, the master control die 304a provides the die scheduling plan via a communication link 608. The communication link 608 provides a path between the master die 304a and the slave die 304b that avoids the use of the memory interface 322. In some embodiments, at least a portion of the communication link 608 extends through the control die TSV 1112. A portion of the communication link 608 may also extend through the memory die TSV 1114.
[0223] Step 2504 includes determining, by the slave control die 304b, whether the die scheduling plan indicates a phase of a programming operation in this time period. In one embodiment, the die scheduling plan indicates the number of upcoming time periods, and whether each time period has been allocated to the memory die 302 controlled by this slave control die 304b. In one embodiment, each of the control die 304 accesses a common clock to track time periods. For example, each control die 304 can count clock cycles in order to convert clock cycles into time periods. If the slave control die 304b is not authorized to perform a phase of a programming operation in this time period, the slave control die 304b waits until a time period allocated to it occurs. When such a time period occurs, in step 2506, the slave control die 304b performs a phase of a programming operation in the allocated time period. Step 2508 includes determining, by the slave control die 304b, whether the programming operation is completed. In step 2508, completion means that the programming loop is no longer performed. If the programming operation is not yet complete, control passes to step 2504 to determine if the next time period has been allocated to control die 304. After the programming operation is complete (step 2508 = Yes), control passes to step 2510. In step 2510, slave control die 304b sends notification of the completion of the programming operation to master control die 304a.
[0224] Since the number of programming loops to be performed is unknown before the master control die 304a issues the die scheduling plan, the master control die 304a may allocate one or more time slots not used by the slave control die 304b to the slave control die 304b. Fig.26 A flow chart depicting one embodiment of a process 2600 for the master control die 304a to update the die schedule in response to receiving notification from the slave control die 304b that it has completed its programming operation. This process 2600 allows for both fine granular control and ensuring that maximum power / current is used in each time period. In other words, the process 2600 can be used to avoid not using the maximum allowed power / current, and therefore not achieving maximum efficiency.
[0225] Step 2602 includes the master control die 304a receiving a notification from the slave control die 304b that the slave control die 304b has completed the programming operation. It should be noted that completing the programming operation means that the programming loop is no longer performed. As with other communications between the master control die 304a and the slave control die 304b, the communication link 608 can be used. Therefore, the memory interface 322 does not need to be used for this communication. Step 2604 includes the master control die 304a selecting another slave control die (or itself) to operate in the previously allocated time period. Step 2606 includes the master control die 304a sending the die schedule to the selected slave control die. Alternatively, the master control die 304a can use the time period to perform memory operations in the memory die 302 it controls.
[0226] In view of the above, it can be seen that the first embodiment includes a device including a stack of semiconductor dies, the stack including multiple control semiconductor dies and multiple memory semiconductor dies. Each semiconductor die in the stack is attached to at least one other semiconductor die in the stack. Each memory semiconductor die includes a memory structure having a non-volatile memory cell, wherein each control semiconductor die is configured to control the operation in the memory structure of at least one of the memory dies. The control semiconductor die includes a master control die and multiple slave control dies. The master control die includes one or more control circuits configured to determine a power plan for performing memory operations in multiple memory semiconductor dies to keep the combined power usage of the stack within a power budget. The power plan includes a slave die scheduling plan for each slave control die and a master die scheduling plan for the master control die. Each die scheduling plan indicates when to perform a memory operation and / or a phase of a memory operation. One or more control circuits of the master control die are further configured to send one of the slave die scheduling plans to each slave control die in the stack. Each respective controlling semiconductor die includes one or more control circuits configured to perform memory operations and / or phases of memory operations at times indicated by a schedule for the respective controlling semiconductor die.
[0227] In a second embodiment, and to facilitate the first embodiment, the one or more control circuits of each control semiconductor die are further configured to determine when to perform memory operations and / or phases of memory operations based on a common clock signal used by all control semiconductor dies in the stack.
[0228] In a third embodiment, and to facilitate either the first or second embodiments, a schedule for each respective slave control die indicates when the respective slave control die is to perform each phase of a programming operation.
[0229] In a fourth embodiment, and to facilitate any of the first to third embodiments, the one or more control circuits of each control semiconductor die are further configured to determine a power profile of a memory semiconductor die controlled by the respective control semiconductor die. Each power profile includes an estimated power usage of the memory semiconductor die in each of a plurality of upcoming time periods. The one or more control circuits of the master control die are further configured to determine a power plan based on the power profile of each memory semiconductor die.
[0230] In a fifth embodiment, and to facilitate any of the first to fourth embodiments, the one or more control circuits of each control semiconductor die are further configured to provide an estimate of the peak current in each phase of a programming operation to be performed. The one or more control circuits of the master control die are further configured to determine a power plan based on the estimate of the peak current.
[0231] In a sixth embodiment, and to facilitate any of the first to fifth embodiments, one or more control circuits of the master control die are further configured to select which memory semiconductor dies to perform memory operations and / or phases of memory operations in a time period to keep the combined power usage of the stack within a power budget during the time period.
[0232] In a seventh embodiment, and to facilitate any of the first to sixth embodiments, one or more control circuits of the master control die are further configured to select which memory semiconductor dies to perform memory operations or phases of memory operations in a time period to maximize the combined power usage of the stack in the time period while keeping the combined power usage of the stack within a power budget.
[0233] In an eighth embodiment, and to facilitate any one of the first to seventh embodiments, one or more control circuits of the master control die are further configured to increase the priority of memory semiconductor dies in the stack to perform memory operations or stages of memory operations in a time period in response to the master control die having prevented the memory semiconductor die from performing memory operations or stages of memory operations in a previous time period.
[0234] In a ninth embodiment, and to facilitate any of the first to eighth embodiments, the memory operation comprises one or more of a programming operation, a read operation, or an erase operation. The phase of the memory operation comprises one or more of a programming phase or a program verification phase of the programming operation.
[0235] In a tenth embodiment, and to facilitate any of the first to ninth embodiments, a plurality of control semiconductor dies each include a through silicon via (TSV) including a communication link, wherein a master control die provides a slave die scheduling plan to a slave control die through the communication link in the TSV of a corresponding slave control die.
[0236] Another embodiment includes a method, which includes analyzing, at a master control die in a stack of semiconductor die, a power usage profile of each of a plurality of memory semiconductor die in the stack in an upcoming time period. The stack includes a memory semiconductor die and a control semiconductor die bonded by a bonding pad. The power usage profile indicates an estimated power usage of each of the memory semiconductor die in each of the upcoming time periods. The method includes developing a plurality of die scheduling plans at a master control die in the stack to remain within the power budget of the stack in each of the time periods. The method includes sending one of the die scheduling plans indicating in which time period a corresponding control semiconductor die performs a phase of a programming operation from the master control die in the stack to each other control semiconductor die. The method includes controlling, by each corresponding control semiconductor die in the stack, the performance of a phase of a programming operation in one of the memory semiconductor die at the indicated time period in the die scheduling plan for the corresponding control semiconductor die.
[0237] Another embodiment includes a nonvolatile storage system, which includes an assembly, the assembly including a plurality of control semiconductor dies and a plurality of memory semiconductor dies bonded by bonding pads, wherein each memory semiconductor die includes a memory structure having a nonvolatile memory cell. Each control semiconductor die is configured to control the operation of the memory structure of at least one of the memory semiconductor dies by sending a signal through the bonding pad. Each control semiconductor die is further configured to estimate the current usage for performing a stage of a programming operation in the memory semiconductor die controlled by the corresponding control semiconductor die in each of a plurality of future time periods. A master control die of the plurality of control semiconductor dies includes one or more control circuits configured to determine a combination of memory semiconductor dies to operate in each of the future time periods to remain within the current usage limit for the assembly based on the estimated value of the current usage of each memory semiconductor die in the assembly. The one or more control circuits of the master control die are further configured to indicate to other control semiconductor dies in the assembly which time period in the future time period each corresponding control semiconductor die performs each stage of the programming operation.
[0238] The foregoing detailed description of the present invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. In view of the above teachings, many modifications and variations are possible. The described embodiments are selected to best explain the principles of the present invention and its practical application, thereby enabling other technical personnel in the field to best utilize the present invention in various embodiments and with various modifications suitable for the intended specific use. It is hoped that the scope of the present invention will be defined by the appended claims.
Claims
1. A device comprising: A stack of semiconductor dies comprising a plurality of control semiconductor dies and a plurality of memory semiconductor dies, wherein each semiconductor die in the stack is attached to at least one other semiconductor die in the stack, wherein each memory semiconductor die includes a memory structure having non-volatile memory cells, wherein each control semiconductor die is configured to control operations in the memory structure of at least one of the memory semiconductor die, wherein the control semiconductor die includes a master control die and a plurality of slave control die; wherein the master control die includes one or more control circuits configured to: determining a power plan for performing memory operations in the plurality of memory semiconductor dice to keep a combined power usage of the stack within a power budget, the power plan comprising a slave die scheduling plan for each slave control die and a master die scheduling plan for the master control die, wherein each die scheduling plan indicates when memory operations and / or phases of memory operations are to be performed, wherein the memory operations include one or more of program operations, read operations, or erase operations, wherein the phases of the memory operations include one or more of a program phase or a program verify phase of a program operation; and sending one of the slave die schedules to each slave control die in the stack; and Wherein each respective controlling semiconductor die comprises one or more control circuits configured to perform the memory operation and / or the phase of the memory operation at a time indicated by the schedule for the respective controlling semiconductor die.
2. The apparatus of claim 1 , wherein the one or more control circuits of each control semiconductor die are further configured to determine when to perform the memory operation and / or the phase of the memory operation based on a common clock signal used by all control semiconductor dies in the stack.
3. The apparatus of claim 1, wherein the schedule for each respective slave control die indicates when the respective slave control die is to perform each phase of a programming operation.
4. The apparatus according to claim 1, wherein: the one or more control circuits of each control semiconductor die are further configured to determine a power profile for a memory semiconductor die controlled by the respective control semiconductor die, wherein each power profile includes an estimated power usage of the memory semiconductor die in each of a plurality of upcoming time periods; and The one or more control circuits of the master control die are further configured to determine the power plan based on the power profile of each memory semiconductor die.
5. The apparatus of claim 1, wherein: The one or more control circuits of each control semiconductor die are further configured to provide an estimate of the peak current in each phase of a programming operation to be performed; and The one or more control circuits of the master control die are further configured to determine the power plan based on the estimated value of the peak current.
6. The apparatus of claim 1 , wherein the one or more control circuits of the master control die are further configured to: Which of the memory semiconductor dies are selected to perform memory operations and / or phases of memory operations during a time period to keep a combined power usage of the stack within the power budget during the time period.
7. The apparatus of claim 1 , wherein the one or more control circuits of the master control die are further configured to: Which of the memory semiconductor dies are selected to perform memory operations or phases of memory operations in a time period to maximize a combined power usage of the stack in the time period while keeping the combined power usage of the stack within the power budget.
8. The apparatus of claim 1 , wherein the one or more control circuits of the master control die are further configured to: Priority of memory semiconductor dies in the stack to perform memory operations or phases of memory operations in a time period is increased in response to the memory semiconductor die having been prevented by the master control die from performing the memory operation or the phase of the memory operation in a previous time period.
9. The apparatus of claim 1 , wherein the plurality of control semiconductor dies each include a through silicon via (TSV) including a communication link, wherein the master control die provides the slave die scheduling plan to the slave control die through the communication link in the TSV of the corresponding slave control die.
10. A method comprising: analyzing, at a master control die in a stack of semiconductor dies, a power usage profile for each of a plurality of memory semiconductor dies in the stack for an upcoming time period, wherein the stack includes the memory semiconductor dies and a control semiconductor die bonded by bond pads, wherein the power usage profile indicates an estimated power usage of each of the memory semiconductor dies for each of the upcoming time periods; developing, at the master control die in the stack, a plurality of die scheduling plans to stay within a power budget of the stack in each of the time periods; sending one of the die schedules indicating in which time period a corresponding control semiconductor die is to perform a phase of a programming operation from the master control die to each other control semiconductor die in the stack; and Performance of the phase of the programming operation in one of the memory semiconductor die at an indicated time period in the die schedule for the respective control semiconductor die is controlled by each respective control semiconductor die in the stack.
11. The method of claim 10, wherein developing the plurality of die scheduling plans comprises developing plans that remain within the power budget of the stack in each of the time periods and maximize power usage in each of the time periods.
12. The method of claim 10, wherein sending one of the die scheduling plans from the master control die to each other control semiconductor die in the stack comprises: The die schedule is transmitted in a communication link in a through silicon via (TSV) in the other controlling semiconductor die.
13. A non-volatile storage system, comprising: an assembly comprising a plurality of control semiconductor dies and a plurality of memory semiconductor dies bonded by bonding pads, wherein each memory semiconductor die comprises a memory structure having nonvolatile memory cells; wherein each control semiconductor die includes means for controlling operation of the memory structure of at least one of the memory semiconductor die by signaling via the bond pads, wherein each control semiconductor die further includes means for estimating current usage for a phase of a programming operation in the memory semiconductor die controlled by the respective control semiconductor die in each of a plurality of future time periods; wherein a master control die of the plurality of control semiconductor die includes control means for determining, based on an estimate of current usage of each memory semiconductor die in the assembly, a combination of the memory semiconductor die to operate in each of the future time periods to remain within a current usage limit for the assembly; and The control device of the master control die is further used to indicate to the other control semiconductor dies in the assembly which time period in the future time period each respective control semiconductor die shall perform each stage of the programming operation.
14. The non-volatile storage system of claim 13, wherein each controlling semiconductor die is configured to access a common clock and perform each phase of the programming operation during an indicated time period based on the common clock.
15. The non-volatile storage system according to claim 13, wherein the control device of the main control die is further configured to: A combination of the memory semiconductor dies is determined to stage respective programming operations in a time period in order to maximize the current usage of the assembly in the time period while remaining within the current usage limit for the assembly.
16. The non-volatile storage system of claim 13, wherein a first one of the control semiconductor die is further configured to send a notification to the master control die when it has completed a programming operation.
17. The non-volatile storage system according to claim 16, wherein the control device of the main control die is further configured to: allocating the time period given to the first control semiconductor die to a second control semiconductor die of the control semiconductor die; and The second controlling semiconductor die is informed of the phase of the programming operation that it will perform in the allocated time period.
18. The non-volatile storage system according to claim 13, wherein the control device of the main control die is further configured to: A priority of a memory semiconductor die in the future time period is increased in response to a phase in which the memory semiconductor die was not allowed to perform a program operation in a previous time period.
19. A non-volatile storage system according to claim 13, wherein the control device of the main control die includes a processing device, and the processing device is programmed to determine the combination of the memory semiconductor die to operate in each of the future time periods to remain within the current usage limit on the assembly based on the estimated current usage of each memory semiconductor die in the assembly.
