摘要:
A memory system is programmed with minimal program disturb and reduced junction and channel leakage during self-boosting. Pre-charging bias signals are applied to word lines adjacent to a selected word line before a program signal is applied to the selected word line and a pass signal is applied to the remaining word lines. The pre-charging bias signals apply a pre-charge to the memory cells. The pre-charging bias signals are chosen to improve the isolation of the memory cells on word lines adjacent to the selected word line, improve self boost efficiency and reduce current leakage to prevent or reduce program disturb and/or programming errors especially in the inhibited memory cells on the selected word line.
摘要:
A method for partial local self-boosting of a memory cell channel is disclosed. As a part of memory cell channel partial local self-boosting, an isolating memory cell located on a source side of a program inhibited memory cell is turned off and a gating memory cell located on a drain side of the program inhibited memory cell is used to pass a pre-charge voltage to the program inhibited memory cell to provide a pre-charge voltage to a channel of the program inhibited memory cell. Moreover, a pre-charge voltage is passed to a buffering memory cell located on the source side of the program inhibited memory cell to provide a pre-charge voltage to a channel of the buffering memory cell and the gating memory cell that is located on the drain side of the program inhibited memory cell is turned off. During programming, a program voltage is applied to the gate of the program inhibited memory cell where a channel voltage of the program inhibited memory cell is raised above a level raised by the pre-charge voltage.
摘要:
A method for partial local self-boosting of a memory cell channel is disclosed. As a part of memory cell channel partial local self-boosting, an isolating memory cell located on a source side of a program inhibited memory cell is turned off and a gating memory cell located on a drain side of the program inhibited memory cell is used to pass a pre-charge voltage to the program inhibited memory cell to provide a pre-charge voltage to a channel of the program inhibited memory cell. Moreover, a pre-charge voltage is passed to a buffering memory cell located on the source side of the program inhibited memory cell to provide a pre-charge voltage to a channel of the buffering memory cell and the gating memory cell that is located on the drain side of the program inhibited memory cell is turned off. During programming, a program voltage is applied to the gate of the program inhibited memory cell where a channel voltage of the program inhibited memory cell is raised above a level raised by the pre-charge voltage.
摘要:
According to one exemplary embodiment, a semiconductor die includes a memory core array situated over a substrate, where the memory core array includes a number of bitlines, where the bitlines can be situated in a first interconnect metal layer in the semiconductor die. The semiconductor die further includes an interconnect structure situated over the memory core array, where the interconnect structure is situated in a second interconnect metal layer in the semiconductor die and situated over each of the bitlines. The interconnect structure can include at least one interconnect line, which can form an angle with respect to the bitlines that can be greater than 0.0 degrees and less than or equal to 90.0 degrees. The interconnect structure can form one of a number of capacitances with each of the bitlines, where each of the capacitances can be substantially equal in value to each other of the capacitances.
摘要:
A semiconductor device includes a first cascode circuit having a first current mirror amplifying a reference current flowing through a data line of a reference cell, and a second current mirror generating a first potential from an amplified reference current; and a second cascode circuit having a third current mirror amplifying a core current flowing through a data line of a core cell, and a transistor receiving a gate voltage corresponding to the amplified reference current and generating a second potential based on a difference between an amplified core cell current and the amplified reference current. Since the second potential is generated by the difference between the core cell current and the reference cell current, the second potential swings in the full range of the ground power supply voltage to the ground potential, and the range of the amplitude of the power supply voltage can be efficiently utilized. Sensing is enabled for a fine current margin.
摘要:
A semiconductor memory device has 2n word lines, a plurality of bit lines, a plurality of nonvolatile memory cells disposed at each intersection of the word lines and the bit lines, a write circuit for writing data to a memory cell located at an intersection of selected ones of the word lines and the bit lines, and a sense amplifier for reading data out of the memory cells. Further, the semiconductor memory device comprises a first unit for simultaneously selecting a block of 2m (n>m) word lines among the 2n word lines, and a second unit for not selecting a block of 2k (m>k) word lines among the 2m word lines. The second unit does not select the block of 2k word lines, and selects a block of 2k word lines prepared outside the 2n word lines when any one of the 2k word lines among the 2m word lines is defective. Consequently, redundant word lines are effectively employed, write and verify operations are stable, and thereby the yield and performance of the semiconductor memory device are improved.
摘要:
A method of charging a data line to a desired voltage level prior to the data line being sensed in a low power memory device by discharging the data line from a voltage level above the desired voltage level to approximately the desired voltage level. By using N-type transistors to discharge the data line to the desired voltage level, the voltage level can be reached faster with cheaper components.
摘要:
A semiconductor memory device has 2n word lines, a plurality of bit lines, a plurality of nonvolatile memory cells disposed at each intersection of the word lines and the bit lines, a write circuit for writing data to a memory cell located at an intersection of selected ones of the word lines and the bit lines, and a sense amplifier for reading data out of the memory cells. Further, the semiconductor memory device comprises a first unit for simultaneously selecting a block of 2m (n>m) word lines among the 2n word lines, and a second unit for not selecting a block of 2k (m>k) word lines among the 2m word lines. The second unit does not select the block of 2k word lines, and selects a block of 2k word lines prepared outside the 2n word lines when any one of the 2k word lines among the 2m word lines is defective. Consequently, redundant word lines are effectively employed, write and verify operations are stable, and thereby the yield and performance of the semiconductor memory device are improved.
摘要:
A pre-amplifier portion of a sense amplifier for a dual bank architecture simultaneous operation flash memory device is provided. The sense pre-amplifier circuit includes two inverting amplifiers, the second inverting amplifier providing a feedback loop for the first inverting amplifier. In addition, special “kicker” circuitry raises the sense pre-amplifier's input signal line to its operating level. The combination of inverting amplifiers, feedback loop and level raising circuitry is configured to provide higher bandwidths for the sense pre-amplifier to accommodate low capacitive loading resulting from a small memory bank. The combination is also configured to provide faster raising of the input signal line to operating level to accommodate high capacitive loading resulting from a large memory bank. The combination is also configured to provide increased signal margins at the output of the sense pre-amplifier.
摘要:
An address buffering and decoding architecture for a multiple bank (or N bank) simultaneous operation flash memory is described. For the duration of a read operation at one bank of the N banks, a write operation can only be performed on any one of the other N-1 banks. For the duration of a write operation at one bank of the N banks, a read operation can only be performed on any one of the other N-1 banks. The address buffering and decoding architecture includes a control logic circuit, an address selection circuit located at each of the N banks, and address buffer circuitry. The control logic circuit is used to generate N read select signals to select one bank of the N banks for a read operation and N write select signals to select another bank of the N banks for a write operation. Each address selection circuit is configured to receive from the control logic circuit a respective one of the N read select signals and a respective one of the N write select signals. The address buffer circuitry is used to simultaneously provide a write address and a read address in order to access core memory cells. Respective first portions of the write and read addresses are provided to the control logic circuit to generate the respective N read select signals and N write select signals. Respective second portions of the write and read addresses are provided to the respective address selection circuit.