Intel Corporation

Micron Technology Incorporated

Robert L Baltar

Mark E Bauer

Ritesh B Trivedi

Sandeep K Guliani

Balaji Srinivasan

Sherif R Sweha

Ravi Annavajjhala

Kevin W Frary

Marc E Landgraf

Steven D Pudar

Suibin Zhang

Dow-Ping D Wong

Balajl Srinivasan

Methods of operating a memory device are useful in managing wear leveling operations. Such methods include receiving an instruction from a host device in communication with the memory device, wherein the instruction comprises a command portion indicating a desire to identify portions of the memory device to be excluded from wear leveling operations and an argument portion comprising information identifying a particular group of one or more blocks of the plurality of blocks; storing the information identifying the particular group of one or more blocks to a non-volatile memory of the memory device as a portion of information identifying blocks to be excluded from wear leveling operations; and performing one or more wear leveling operations only on a subset of the plurality of blocks responsive to the information identifying blocks to be excluded from wear leveling operation.

Wear leveling for a memory device

Memory devices and methods to facilitate wear leveling operations in a memory device. In one such method, particular blocks of memory cells are excluded from experiencing wear leveling operations performed on the memory device. In at least one method, a user selects blocks of memory to be excluded from wear leveling operations performed on the remainder of blocks of the memory device. Selected blocks of memory are excluded from wear leveling operations responsive to a command initiated by a user identifying, either directly or indirectly, the selected blocks to be excluded.

An apparatus and method are disclosed for providing a kicker function for non-volatile memory drain bias. According to one embodiment, the kicker function is provided by a high performance transistor that is activated by a kicker enable signal, providing a kicker function for non-volatile memory drain bias. According to one embodiment, the kicker function is provided by a high performance transistor that is activated by a kicker enable signal.

Kicker for non-volatile memory drain bias

An apparatus for a differential redundancy multiplexor for flash memory devices. One embodiment comprises a memory array comprising a main memory element and a redundant element. A sense amp is coupled to the memory array to evaluate the main memory element and to generate a first pair of differential output signals. A redundant sense amp is coupled to the memory array. The redundant sense amp is to evaluate the redundant memory element and to generate a second pair of differential output signals. A multiplexor is coupled to the sense amp and the redundant sense amp. The multiplexor is to receive the first pair and the second pair. The multiplexor is to generate a single ended output from evaluating a single pair of differential output signals. Control logic coupled to the multiplexor to control whether the first pair or the second pair is the single pair of differential output signals evaluated.

Differential redundancy multiplexor for flash memory devices

An apparatus is disclosed for providing a load for a non-volatile memory drain bias circuit. Under an embodiment, a load for a non-volatile memory drain bias circuit comprises a column load and a current mirror, a reference voltage for the current mirror being a sample and hold voltage reference. The column load and the current mirror are coupled to a cascode device.

Load for non-volatile memory drain bias

An apparatus and method are disclosed for providing drain bias for non-volatile memory. According to one embodiment, the drain bias is provided utilizing a drain bias circuit that is referenced by a static voltage reference.

Drain bias for non-volatile memory

A method and apparatus for a memory device is described. In one embodiment, global Y (GY) enable is gated by the trailing edge of a address transition detection (ATD) pulse. The ATD pulse ensures that the GY enable is off during periods when the memory device is not attempting to read a memory cell. The sense (SEN) node between the GY transistor and drain bias circuit may be charged up and global bit line (GBL) may be grounded. During this time, the power supply current is cut off by the GY transistor itself, thereby eliminating the need of separate cut-off transistors within the drain bias circuit. This permits minimal time delay in sensing after the incoming address is stable.

Method and apparatus for gating a global column select line with address transition detection

An apparatus and method for protecting memory blocks in a block-based flash erasable programmable read only memory device are disclosed. A non-volatile memory array includes a number of blocks that are capable of being placed in a locked state or an unlocked state. A volatile lock register is coupled to each of the lockable blocks in the memory array. A logic state is coupled to one input of the volatile lock register, and a block set/reset line is coupled to a second input of the volatile lock register. A block latch control line is coupled to one input of the logic gate, and a group latch control line is coupled to a second input of the logic gate. The method includes reading a first command of a multi-cycle command specifying a lock configuration of one or more memory blocks and reading a second command specifying the number of memory blocks to be lock configured. The method and apparatus allow a user to dynamically select which blocks of a flash array to lock or unlock and minimize the possibility of data corruption during block lock and unlock cycles.

Computing system with volatile lock architecture for individual block locking on flash memory

A flash memory using a pre-sensing amplifier coupled to receive differential inputs from a pair of memory cells of said flash memory array and to generate a differential output from the pre-sensing amplifier. The differential output is coupled to a bus, which is also coupled to a post-sensing amplifier. The differential configuration on the bus allows marginal voltage differences to be detected by the post-sensing amplifier so that logic states from the flash memory can be sensed without the bus transitioning to half of the supply voltage.

Differential signal path for high speed data transmission in flash memory

An apparatus and method are disclosed for providing a sample and hold voltage reference for non-volatile memory. According to one embodiment, the sample and hold voltage reference produces a reference voltage for a drain bias circuit of a memory cell.

Sample and hold voltage reference source

An embodiment of the invention is directed to a method of operating a flash memory, which includes discharging at least one local wordline of an unselected block of flash memory cells during an interval in which a selected set of flash memory cells are being conditioned, such that the at least one local wordline does not develop a charge that is sufficient to corrupt the data stored in the unselected block.

Preventing data corruption in a memory device using a modified memory cell conditioning methodology

An apparatus and method for protecting memory blocks in a block-based flash erasable programmable read only memory (EPROM) device are disclosed. A non-volatile memory array includes a number of blocks that are capable of being placed in a locked state or an unlocked state. A volatile lock register is coupled to each of the lockable blocks in the memory array. A logic gate is coupled to one input of the volatile lock register, and a block set/reset line is coupled to a second input of the volatile lock register. A block latch control line is coupled to one input of the logic gate, and a group latch control line is coupled to a second input of the logic gate. The method includes reading a first command of a multi-cycle command specifying a lock configuration of one or more memory blocks and reading a second command specifying the number of memory blocks to be lock configured. The method and apparatus allow a user to dynamically select which blocks of a flash array to lock or unlock and minimize the possibility of data corruption during block lock and unlock cycles.

Method and apparatus using volatile lock architecture for individual block locking on flash memory

A nonvolatile memory includes a global line and a first block and a second block. The first block includes a plurality of first local lines and a first local decoder coupled to the global line and the first local lines for selectively coupling the global line to one of the first local lines in accordance with an address when the first local decoder is enabled and for isolating the first local lines from the global line when the first local decoder is disabled. The second block includes a plurality of second local lines and a second local decoder coupled to the global line and the second local lines for selectively coupling the global line to one of the second local lines in accordance with the address when the second local decoder is enabled and for isolating the second local lines from the global line when the second local decoder is disabled such that interference between the first and second blocks is eliminated during memory operations.

Nonvolatile memory blocking architecture

A method and circuitry for providing dual row selection using a multiplexed tri-level decoder is disclosed. For one embodiment, the multiplexed tri-level decoder is a 3:8 decoder, the major components of which are a buffer and 8 three input NAND circuits. The NAND circuits are peculiar in that the inputs are referenced to a VCC operational voltage supply, and the outputs are referenced to a VPX tri-level supply voltage. The output of each NAND circuit is used to select one row or word line. During preconditioning and post conditioning, the decoder is required to enable two adjacent rows: the row selected and the next row. The present design implements dual row selection by adding a pass transistor that connects the word line enable driver to the driver of the previous row within the VPX level circuitry. This is in contrast to the previous design approach of implementing dual row selection by using VCC level logic. The disclosed implementation eliminates gates in the speed path of the circuit. This reduces the delay through the circuit. The new generation of flash memory requires reduction in read access time compared to that of previous generations. The disclosed tri-level decoder eliminates gates in the critical speed path, directly reducing read access time for the memory array.

Growing community of inventors

Folsom, CA, United States of America

Robert L Baltar