Abstract: A data storage device includes a power supply circuit configured to supply power to the data storage device. The power supply circuit includes a voltage clamp configured to operate in a conduction state in response to an over-voltage condition of the power supply circuit. The power supply circuit also includes a fuse in series with the voltage clamp. The fuse is configured to open in response to a current flow through the fuse and the voltage clamp exceeding a threshold value. The power supply circuit also includes a switching device that is configured to latch in a forward conduction mode in response to the voltage clamp operating in the conduction state. The switching device couples power from a positive voltage bus to the voltage clamp when the switching device is in the forward conduction mode.

Abstract: Latch-based multiply-accumulate (MAC) operations implemented on the die of a non-volatile memory (NVM) array are disclosed. The exemplary latch-based MAC procedures described herein are linear procedures that do not require logic branches. In one example, the MAC operation uses a set of linear MAC stages, wherein each linear stage processes MAC operations corresponding to one bit of a first multi-bit multiplicand being multiplied against a second multi-bit multiplicand. Examples are provided wherein the MAC procedures are performed as part of a neural network feedforward procedure where the first multiplicand is a synaptic weight and the second multiplicand is an activation value. Multiple plane and multiple die NVM array implementations are also described for massive parallel processing.

Abstract: Noise injection procedures implemented on the die of a non-volatile memory (NVM) array are disclosed. In one example, noise is injected into data by adjusting read voltages to induce bit flips while using feedback to achieve a target amount of information degradation. In another example, random data is iteratively combined with itself to achieve a target percentage of random 1s or 0s, then the random data is combined with data read from the NVM array. In other examples, pixels are randomly zeroed out to emulate dead charge coupled device (CCD) pixels. In still other examples, the timing, voltage, and/or current values used within circuits while transferring data to/from latches or bitlines are adjusted outside their specified margins to induce bit flips to inject noise into the data. The noise-injected data may be used, for example, for dataset augmentation or for the testing of deep neural networks (DNNs).

Abstract: Latch-based methods and apparatus for performing neural network weight parity detection on the die of a non-volatile memory (NVM) array to detect bit flip errors within neural network weight data are described, particularly for use with floating point number values. Upon detection of a parity error in a neural network weight, the erroneous weight is set to zero to trim the corresponding neuron from the network, thus preventing the erroneous value from significantly affecting the network, particularly in situations where the bit flip would otherwise affect the magnitude of a floating-point weight value. The exemplary latch-based procedures described herein are linear procedures that do not require logic decisions. Procedures are also described that assess an amount of degradation in the NVM array based on parity bit data collected in the latches. Multiple plane and multiple die NVM array implementations are also described for massive parallel processing.

Abstract: Noise injection procedures implemented on the die of a non-volatile memory (NVM) array are disclosed. In one example, noise is injected into data by adjusting read voltages to induce bit flips while using feedback to achieve a target amount of information degradation. In another example, random data is iteratively combined with itself to achieve a target percentage of random 1s or 0s, then the random data is combined with data read from the NVM array. In other examples, pixels are randomly zeroed out to emulate dead charge coupled device (CCD) pixels. In still other examples, the timing, voltage, and/or current values used within circuits while transferring data to/from latches or bitlines are adjusted outside their specified margins to induce bit flips to inject noise into the data. The noise-injected data may be used, for example, for dataset augmentation or for the testing of deep neural networks (DNNs).

Abstract: Bit-flip object insertion techniques are provided for use with a non-volatile memory (NVM) wherein an object is inserted into a background image by flipping or inverting one or more bits within the pixels of the background image that correspond to the shape and insertion location of an object being inserted. In an illustrative example, pixels within the background image that correspond to the shape and insertion location of the object are XORed with binary 1s. This flips the bits of those pixels to change the color (hue) and/or intensity (brightness) of the pixels so the object appears in the background image. In other examples, only the most significant bits of pixels in the background image are inverted (flipped). Exemplary latch-based procedures are described herein for high-speed processing on an NVM die. Multiple plane NVM die implementations are also described for massive processing.

Abstract: Embodiments of the present disclosure generally include methods of specially programming a set of memory cells, wherein each specially programmed memory cell is specially programmed along with programming a plurality of wordlines, and wherein each memory cell is specially programmed by altering a bitline and gate voltage applied to the memory cell. The methods further includes performing a sensing operation across a set of strings in the array of memory cells, determining, based on the sensing operation, whether one or more strings failed to conduct during a sensing operation, and determining the last programmed wordline using the one or more strings that failed to conduct.

Abstract: A data storage device includes a non-volatile memory device including a memory block having a number of memory dies, and a controller coupled to the memory device. A memory access command is received and a memory access operation based on the received command is performed. A number of bytes transferred during the memory access operation is determined, and the determined number of bytes is analyzed to determine whether the number of transferred bytes is equal to a predetermined number. A transfer status fail bit is set if the number of transferred bytes is not equal to the predetermined number.

Abstract: Clamped semiconductor wafers and clamped semiconductor devices include reservoirs filled with a flowable metal which hardens to allow the wafers/devices to be shipped or stored. The hardened metal may also be reflowed to a liquid to allow clamping of the semiconductor wafers together and to allow clamping of the semiconductor packages together. The flowable metal may be filled into the reservoirs as a liquid or paste. Thereafter, the flowable metal may be cooled to harden the flowable metal into a clamping member.

Abstract: A die-to-die voltage sharing process that may be implemented to overcome a charge pump failure on a memory die of a non-volatile storage device. When a charge pump failure is detected, a controller causes another memory die with a functional charge pump to generate and supply a voltage to the memory die with the failed charge pump. When the voltage is received by the memory die with the failed charge pump, the voltage may be used to perform a requested memory operation.

Abstract: A die-to-die voltage sharing process that may be implemented to overcome a charge pump failure on a memory die of a non-volatile storage device. When a charge pump failure is detected, a controller causes another memory die with a functional charge pump to generate and supply a voltage to the memory die with the failed charge pump. When the voltage is received by the memory die with the failed charge pump, the voltage may be used to perform a requested memory operation.

Abstract: A data storage device includes a power supply circuit configured to supply power to the data storage device. The power supply circuit includes a voltage clamp configured to operate in a conduction state in response to an over-voltage condition of the power supply circuit. The power supply circuit also includes a fuse in series with the voltage clamp. The fuse is configured to open in response to a current flow through the fuse and the voltage clamp exceeding a threshold value. The power supply circuit also includes a switching device that is configured to latch in a forward conduction mode in response to the voltage clamp operating in the conduction state. The switching device couples power from a positive voltage bus to the voltage clamp when the switching device is in the forward conduction mode.

Abstract: A system and method for warpage detection in a CMOS bonded array includes a conductor positioned between bonded contact pads of first and second wafers. The conductor is connected to a continuity check circuit. If the continuity check circuit detects an interruption in conductivity of the conductor, such interruption is indicative of warpage in the first and/or second wafers. In one implementation, the conductor is a serpentine-shaped structure.

Abstract: A data storage device includes a memory device, an internal power supply, and a controller. When the data storage device is in an idle power state, the controller is configured to electrically isolate the data storage device from an external power source and draw power from the internal power supply to perform data storage operations. The power utilized for the data storage operations while the data storage device is in the idle power state may be exclusively from the internal power supply or a mixture of the internal power supply and the external power supply. The total power utilized during the idle power state is less than a threshold that is relative to the active power of the data storage device or at a fixed value.

Abstract: A data storage device includes a non-volatile memory device including a memory block having a number of memory dies, and a controller coupled to the memory device. A memory access command is received and a memory access operation based on the received command is performed. A number of bytes transferred during the memory access operation is determined, and the determined number of bytes is analyzed to determine whether the number of transferred bytes is equal to a predetermined number. A transfer status fail bit is set if the number of transferred bytes is not equal to the predetermined number.

Abstract: A non-volatile storage system includes a control circuit connected to non-volatile memory cells provides for progressive writing of data. That is, existing data is overwritten by new data without performing a traditional erase operation that changes the threshold voltage of the memory cells back to the traditional or original erase state. In one example, new data is written on top of old data using shifted threshold voltage distributions. Some embodiments include writing MLC data over SLC data, using intermediate erase threshold voltage distributions and/or automatically detecting which threshold voltage distributions are currently being used to store data.

Abstract: A semiconductor device including an integrated module formed of a first semiconductor die coupled to a second semiconductor die. Each of the first and second semiconductor dies includes a number of bond pads, which are bonded to each other to form the integrated module. Each bond pad may be divided into a number of discrete pad legs. While the overall footprint of each bond pad on the first and second semiconductor dies may be the same, the bond pads on one of the dies may have a larger number of pad legs.

Abstract: A method including stacking a number of silicon dice such that one or more edges of the dice are in vertical alignment, where the one or more edges include a number of connection pads. The method also includes positioning a connecting wire on a substantially perpendicular axis to the one or more edges. The connecting wire includes a number of solder blocks formed thereon. The solder blocks are spaced at intervals associated with a distance between a first set of aligned connection pads on the dice. The connecting wire is positioned such that the solder blocks are in contact with the first set of aligned connection pads. The method also includes applying heat to cause the solder blocks to reflow and physically and electrically couple the connecting wire to the connection pads.

Abstract: A data storage device includes a memory device, an internal power supply, and a controller. When the data storage device is in an idle power state, the controller is configured to electrically isolate the data storage device from an external power source and draw power from the internal power supply to perform data storage operations. The power utilized for the data storage operations while the data storage device is in the idle power state may be exclusively from the internal power supply or a mixture of the internal power supply and the external power supply. The total power utilized during the idle power state is less than a threshold that is relative to the active power of the data storage device or at a fixed value.

Abstract: A system and method for die crack detection in a CMOS bonded array includes a capacitor that is formed in an edge seal, where one of the capacitor plates is in the edge seal portion of a first wafer, and the other capacitor plate is in the edge seal portion of a second wafer. A crack in the die can be detected by applying an alternating current waveform to a testing contact on the integrated circuit and monitoring for a shift in the alternating current waveform, as a crack would cause modulation in the capacitance.