Abstract: A three-dimensional array of memory elements is formed across multiple layers of planes positioned at different distances above a semiconductor substrate. The memory elements reversibly change a level of electrical conductance in response to a voltage difference being applied across them. The three-dimensional array includes a two-dimensional array of pillar lines from the substrate through the multiple layers of planes. A first set of pillar lines acts as local bit lines for accessing the memory elements together with an array of word lines on each plane. A second set of pillar lines is connected to the word lines. An array of metal lines on the substrate is switchable connected to the pillar lines to provide access to the first and second sets of pillar lines, thereby to provide access respectively to the bit lines and word lines of the three-dimensional array.

Abstract: A non-volatile memory core comprises one or more memory bays. Each memory bay comprises one or more memory blocks that include a grouping of non-volatile storage elements. In one embodiment, memory blocks in a particular memory bay share a group of read/write circuits. During a memory operation, memory blocks are transitioned into active and inactive states. The process of transitioning blocks from an inactive state to an active state includes enabling charge sharing between a memory block entering the active state and another memory block that was previously in the active state. This charge sharing improves performance and/or reduces energy consumption for the memory system.

Abstract: A non-volatile storage system connects a signal driver to a first control line that is connected to a first non-volatile storage element, charges the first control line using the signal driver while the signal driver is connected to the first control line, disconnects the signal driver from the first control line while the first control line remains charged from the signal driver, connects the signal driver to a second control line that is connected to a second non-volatile storage element, charges the second control line using the signal driver while the signal driver is connected to the second control line, and disconnects the signal driver from the second control line. Charging the control lines causes the respective non-volatile storage elements to experience a program operation.

Abstract: A technique for identifying bad pages of storage elements in a memory device. A flag byte is provided for each page group of one or more pages which indicates whether the page group is healthy. Flag bytes of selected page groups also indicate whether larger sets of page groups are healthy, according to bit positions in the flag bytes. A bad page identification process includes reading the flag bytes with a selected granularity so that not all flag bytes are read. Optionally, a drill down process reads flag bytes for smaller sets of page groups when a larger set of page groups is identified as having at least one bad page. This allows the bad page groups to be identified and marked with greater specificity. Redundant copies of flag bytes may be stored in different locations of the memory device. A majority vote process assigns a value to each bit.

Abstract: A storage system includes a three-dimensional memory array that has multiple layers of non-volatile storage elements grouped into blocks. The blocks are grouped into bays. The storage system includes array lines of a first type in communication with storage elements, array lines of a second type in communication with storage elements, and sense amplifiers. Each block is geographically associated with two sense amplifiers and all blocks of a particular bay share a group of sense amplifiers associated with the blocks of the particular bay. The system includes multiple sets of local data lines in one or more routing metal layers below the three-dimensional memory array and multiple sets of global data lines in one or more top metal layers above the three-dimensional memory array. Each set of one or more blocks include one set of the local data lines. Each bay includes one set of global data lines that connect to the group of sense amplifiers associated with the blocks of the respective bay.

Abstract: The embodiments described herein are directed to a memory device with multi-level, write-once memory cells. In one embodiment, a memory device has a memory array comprising a plurality of multi-level write-once memory cells, wherein each memory cell is programmable to one of a plurality of resistivity levels. The memory device also contains circuitry configured to select a group of memory cells from the memory array, and read a set of flag bits associated with the group of memory cells. The set of flag bits indicate a number of times the group of memory cells has been written to. The circuitry is also configured to select a threshold read level appropriate for the number of times the group of memory cells has been written to, and for each memory cell in the group, read the memory cell as an unprogrammed single-bit memory cell or as a programmed single-bit memory cell based on the selected threshold read level.

Abstract: Column redundancy data is selectively retrieved in a memory device according to a set of storage elements which is currently being accessed, such as in a read or write operation. The memory device is organized into sets of storage elements such as logical blocks, where column redundancy data is loaded from a non-volatile storage location to a volatile storage location for one or more particular blocks which are being accessed. The volatile storage location need only be large enough to store the current data entries. The size of the set of storage elements for which column redundancy data is concurrently loaded can be configured based on an expected maximum number of defects and a desired repair probability. During a manufacturing lifecycle, the size of the set can be increased as the number of defects is reduced due to improvements in manufacturing processes and materials.

Abstract: A non-volatile storage system includes technology for skipping programming cycles while programming a page (or other unit) of data. While programming a current subset of the page (or other unit) of data, the system will evaluate whether the next subsets of the page (or other unit) of data should be programmed into non-volatile storage elements or skipped. Subsets of the page (or other unit) of data that should not be skipped are programmed into non-volatile storage elements. Some embodiments include transferring the appropriate data to temporary latches/registers, in preparation for programming, concurrently with the evaluation of whether to program or skip the programming.

Abstract: A non-volatile storage device includes a substrate, a monolithic three-dimensional memory array of non-volatile storage elements arranged above a portion of the substrate, a plurality of sense amplifiers in communication with the non-volatile storage elements, a plurality of temporary storage devices in communication with the sense amplifiers, a page register in communication with the temporary storage devices, and one or more control circuits. The one or more control circuits are in communication with the page register, the temporary storage devices and the sense amplifiers. The sense amplifiers are arranged on the substrate underneath the monolithic three-dimensional memory array. The temporary storage devices are arranged on the substrate underneath the monolithic three-dimensional memory array. The page register is arranged on the substrate in an area that is not underneath the monolithic three-dimensional memory array.

Abstract: A method of biasing a nonvolatile memory array. The nonvolatile memory array includes a first and second plurality of Y lines, a plurality of X lines, a first and second plurality of two terminal memory cells. Each first and second memory cell is coupled to one of the first or second plurality of Y lines and one of the plurality of X lines, respectively. Substantially all of the first plurality and second plurality of Y lines are driven to a Y line unselect voltage. At least one selected Y line of the first plurality of Y lines is driven to a Y line select voltage while floating remaining Y lines of the first plurality of Y lines and while driving substantially all of the second plurality of Y lines to the Y line unselect voltage.

Abstract: A non-volatile memory core comprises one or more memory bays. Each memory bay comprises one or more memory blocks that include a grouping of non-volatile storage elements. In one embodiment, memory blocks in a particular memory bay share a group of read/write circuits. During a memory operation, memory blocks are transitioned into active and inactive states. The process of transitioning blocks from an inactive state to an active state includes enabling charge sharing between a memory block entering the active state and another memory block that was previously in the active state. This charge sharing improves performance and/or reduces energy consumption for the memory system.

Abstract: A memory system includes a substrate, control circuitry on the substrate, a three dimensional memory array (above the substrate) that includes a plurality of memory cells with reversible resistance-switching elements, and a circuit for detecting the setting and resetting of the reversible resistance-switching elements. In one aspect a circuit that has one or more clock inputs is run for a predetermined number of clock cycles. The circuit generates an amount of charge over the predetermined number of clock cycles. At most the amount of charge is provided to non-volatile storage element to program the non-volatile storage element. It is determined whether the non-volatile storage element is programmed to a desired state as a result of providing at most the amount of charge to the non-volatile storage element. Techniques disclosed herein can be applied to program memory cells other than memory cells with reversible resistance-switching elements.

Abstract: A memory system includes a substrate, control circuitry on the substrate, a three dimensional memory array (above the substrate) that includes a plurality of memory cells with reversible resistance-switching elements, and a circuit for detecting the setting and resetting of the reversible resistance-switching elements. In one aspect a circuit that has one or more clock inputs is run for a predetermined number of clock cycles. The circuit generates an amount of charge over the predetermined number of clock cycles. At most the amount of charge is provided to non-volatile storage element to program the non-volatile storage element. It is determined whether the non-volatile storage element is programmed to a desired state as a result of providing at most the amount of charge to the non-volatile storage element. Techniques disclosed herein can be applied to program memory cells other than memory cells with reversible resistance-switching elements.

Abstract: A method of biasing a nonvolatile memory array. The nonvolatile memory array includes a first and second plurality of Y lines, a plurality of X lines, a first and second plurality of two terminal memory cells. Each first and second memory cell is coupled to one of the first or second plurality of Y lines and one of the plurality of X lines, respectively. Substantially all of the first plurality and second plurality of Y lines are driven to a Y line unselect voltage. At least one selected Y line of the first plurality of Y lines is driven to a Y line select voltage while floating remaining Y lines of the first plurality of Y lines and while driving substantially all of the second plurality of Y lines to the Y line unselect voltage.

Abstract: A storage system includes a three-dimensional memory array that has multiple layers of non-volatile storage elements grouped into blocks. Each block includes a subset of first selection circuits for selectively coupling a subset of array lines (e.g. bit lines) of a first type to respective local data lines. Each block includes a subset of second selection circuits for selectively coupling a subset of the respective local data lines to global data lines that are connected to control circuitry. To increase the performance of memory operations, the second selections circuits can change their selections independently of each other. For example, a memory operation is performed concurrently on a first non-volatile storage element of each group of a plurality of groups of non-volatile storage elements. Completion of the memory operation for the first non-volatile storage element of each group is independently detected.

Abstract: A three-dimensional array especially adapted for memory elements that reversibly change a level of electrical conductance in response to a voltage difference being applied across them. Memory elements are formed across a plurality of planes positioned different distances above a semiconductor substrate. A two-dimensional array of bit lines to which the memory elements of all planes are connected is oriented vertically from the substrate and through the plurality of planes. A double-global-bit-line architecture provides a pair of global bit lines for each bit lines for accessing a row of memory elements in parallel. A first one of each pair allows the local bit lines of the row to be sensed while a second one of each pair allows local bit lines in an adjacent row to be set to a definite voltage so as to eliminate leakage currents between adjacent rows of local bit lines.

Abstract: A cross point memory cell includes a portion of a first distributed diode, a portion of a second distributed diode, a memory layer located between the portion of the first distributed diode and the portion of a second distributed diode, a bit line electrically connected to the first distributed diode, and a word line electrically connected to the second distributed diode.

Abstract: A technique for identifying bad pages of storage elements in a memory device. A flag byte is provided for each page group of one or more pages which indicates whether the page group is healthy. Flag bytes of selected page groups also indicate whether larger sets of page groups are healthy, according to bit positions in the flag bytes. A bad page identification process includes reading the flag bytes with a selected granularity so that not all flag bytes are read. Optionally, a drill down process reads flag bytes for smaller sets of page groups when a larger set of page groups is identified as having at least one bad page. This allows the bad page groups to be identified and marked with greater specificity. Redundant copies of flag bytes may be stored in different locations of the memory device. A majority vote process assigns a value to each bit.

Abstract: A technique for efficiently handling write operation failures in a memory device which communicates with an external host device allows a page of data to be re-written to a memory array from a page buffer. The host provides user data, a first write address and a write command to the memory device. If the write attempt fails, the host provides a re-write command with a new address, without re-sending the user data to the memory device. Additional data can be received at a data cache of the memory device while a re-write from the page buffer is in progress. The re-written data may be obtained in a copy operation in which the data is read out to the host, modified and written back to the memory device. Additional data can be input to the memory device during the copy operation. Page buffer data can also be modified in place.

Abstract: Column redundancy data is selectively retrieved in a memory device according to a set of storage elements which is currently being accessed, such as in a read or write operation. The memory device is organized into sets of storage elements such as logical blocks, where column redundancy data is loaded from a non-volatile storage location to a volatile storage location for one or more particular blocks which are being accessed. The volatile storage location need only be large enough to store the current data entries. The size of the set of storage elements for which column redundancy data is concurrently loaded can be configured based on an expected maximum number of defects and a desired repair probability. During a manufacturing lifecycle, the size of the set can be increased as the number of defects is reduced due to improvements in manufacturing processes and materials.