Abstract: Provided is a method of forming a trench gate MOSFET. A hard mask layer is formed on a substrate. The substrate is partially removed by using the hard mask layer as a mask, so as to form a trench in the substrate. A first insulating layer and a first conductive layer are formed in the lower portion of the trench. A sacrificial layer is formed on the side surface of the upper portion of the trench, and the sacrificial layer is connected to the hard mask layer. An interlayer insulating layer is formed on the first conductive layer by a thermal oxidation process when the sacrificial layer and the hard mask layer are present. A second insulating layer and a second conductive layer are formed in the upper portion of the trench. A trench gate MOSFET is further provided.

Abstract: Provided is a method of forming a trench gate MOSFET. A hard mask layer is formed on a substrate. The substrate is partially removed by using the hard mask layer as a mask, so as to form a trench in the substrate. A first insulating layer and a first conductive layer are formed in the lower portion of the trench. A sacrificial layer is formed on the side surface of the upper portion of the trench, and the sacrificial layer is connected to the hard mask layer. An interlayer insulating layer is formed on the first conductive layer by a thermal oxidation process when the sacrificial layer and the hard mask layer are present. A second insulating layer and a second conductive layer are formed in the upper portion of the trench. A trench gate MOSFET is further provided.

Abstract: Provided is a method of forming a trench gate MOSFET. A hard mask layer is formed on a substrate. The substrate is partially removed by using the hard mask layer as a mask, so as to form a trench in the substrate. A first insulating layer and a first conductive layer are formed in the lower portion of the trench. A sacrificial layer is formed on the side surface of the upper portion of the trench, and the sacrificial layer is connected to the hard mask layer. An interlayer insulating layer is formed on the first conductive layer by a thermal oxidation process when the sacrificial layer and the hard mask layer are present. A second insulating layer and a second conductive layer are formed in the upper portion of the trench. A trench gate MOSFET is further provided.

Abstract: Provided is a method of forming a trench gate MOSFET. A hard mask layer is formed on a substrate. The substrate is partially removed by using the hard mask layer as a mask, so as to form a trench in the substrate. A first insulating layer and a first conductive layer are formed in the lower portion of the trench. A sacrificial layer is formed on the side surface of the upper portion of the trench, and the sacrificial layer is connected to the hard mask layer. An interlayer insulating layer is formed on the first conductive layer by a thermal oxidation process when the sacrificial layer and the hard mask layer are present. A second insulating layer and a second conductive layer are formed in the upper portion of the trench. A trench gate MOSFET is further provided.

Abstract: A method of sorting a multi-bit per cell non-volatile memory includes programming and reading to test an n-bit-per-cell (n-bpc) non-volatile memory, which has a plurality of m-bpc pages, where m is a positive integer from 1 through n. If the m-bpc page fails the test, counting a block associated with the failed m-bpc page to (m-1)-bpc blocks, wherein each said m-bpc page is subjected to at most one time of programming and reading. When m is equal to 1, the 0-bpc block corresponds to a bad block.

Abstract: A method of programming a multi-bit per cell non-volatile memory is disclosed. In one embodiment, the non-volatile memory is read to obtain a first data of a most-significant-bit (MSB) page on a current word line that succeeds in data reading, wherein the current word line follows a preceding word line on which data reading fails. At least one reference voltage is set. The MSB page on the current word line is secondly programmed with a second data according to the reference voltage, the second data being different from the first data.

Abstract: A method of accessing a non-volatile memory is disclosed. Original bits of data are duplicated on a bit level to generate a plurality of duplicated bits corresponding to each original bit. At least one shielding bit is provided between the duplicated bits corresponding to different original bits. The duplicated bits and the at least one shielding bit are programmed to the non-volatile memory. The original bits are generated or determined according to the duplicated bits.

Abstract: A method of programming/reading a multi-bit per cell non-volatile memory with a sequence is disclosed. A plurality of less-significant-bit pages are programmed, and a plurality of consecutive most-significant-bit pages of a plurality of consecutive word lines are programmed one after the other in a consecutive order. The most-significant-bit pages through all word lines in at least one memory block of the non-volatile memory are programmed or read after the less-significant-bit pages through all the word lines in the at least one memory block are programmed or read.

Abstract: A method of dynamic data storage for error correction in a memory device is disclosed. Data for storage is received, the received data is encoded and error correction code (ECC) is generated. The encoded data is stored in the memory device that includes a plurality of pages each having a plurality of data partitions. More corrected errors a marked page has, a smaller portion with a space of at least one datum of each of the corresponding data partitions associated with the marked page is allocated to store the encoded data, while a size of the ECC is fixed, thereby increasing capability of correcting errors in the marked page.

Abstract: A method of accessing a non-volatile memory is disclosed. Original bits of data are duplicated on a bit level to generate a plurality of duplicated bits corresponding to each original bit. At least one shielding bit is provided between the duplicated bits corresponding to different original bits. The duplicated bits and the at least one shielding bit are programmed to the non-volatile memory. The original bits are generated or determined according to the duplicated bits.

Abstract: A method of programming/reading a multi-bit per cell non-volatile memory with a sequence is disclosed. A plurality of less-significant-bit pages are programmed, and a plurality of consecutive most-significant-bit pages of a plurality of consecutive word lines are programmed one after the other in a consecutive order. The most-significant-bit pages through all word lines in at least one memory block of the non-volatile memory are programmed or read after the less-significant-bit pages through all the word lines in the at least one memory block are programmed or read.

Abstract: A method of programming a multi-bit per cell non-volatile memory is disclosed. In one embodiment, the non-volatile memory is read to obtain a first data of a most-significant-bit (MSB) page on a current word line that succeeds in data reading, wherein the current word line follows a preceding word line on which data reading fails. At least one reference voltage is set. The MSB page on the current word line is secondly programmed with a second data according to the reference voltage, the second data being different from the first data.

Abstract: A method of twice programming a multi-bit per cell non-volatile memory with a sequence is disclosed. At least one page at a given word line is firstly programmed with program data by a controller of the non-volatile memory, and at least one page at a word line preceding the given word line is secondly programmed with the same program data by the controller.

Abstract: A method of programming a multi-bit per cell non-volatile memory is disclosed. In one embodiment, the non-volatile memory is read to obtain a first data of a most-significant-bit (MSB) page on a current word line that succeeds in data reading, wherein the current word line follows a preceding word line on which data reading fails. At least one reference voltage is set. The MSB page on the current word line is secondly programmed with a second data according to the reference voltage, the second data being different from the first data.

Abstract: A method of dynamic data storage for error correction in a memory device is disclosed. Data for storage is received, and the received data is then encoded and associated error correction code (ECC) is generated. The encoded data is stored in a portion of a data partition of the memory device, wherein percentage of the stored data in the data partition is determined according to an amount of corrected errors associated with the data partition or is predetermined.

Abstract: A method of sorting a multi-bit per cell non-volatile memory includes programming and reading to test an n-bit-per-cell (n-bpc) non-volatile memory, which has a plurality of m-bpc pages, where m is a positive integer from 1 through n. If the m-bpc page fails the test, counting a block associated with the failed m-bpc page to (m-1)-bpc blocks, wherein each said m-bpc page is subjected to at most one time of programming and reading. When m is equal to 1, the 0-bpc block corresponds to a bad block.

Abstract: A program method for a non-volatile memory is disclosed. At least two blocks in the non-volatile memory are configured as 1-bit per cell (1-bpc) blocks. The data of the configured blocks are read and written to a target block in such a way that the data of each said configured block are moved to pages of a same significant bit. In another embodiment, the data of the configured blocks excluding one block are read and written to the excluded block.

Abstract: A frequency counter obtains a cycle number of a clock of a target signal by a reference signal and a clock mask synchronous with the target signal, calculates a frequency of the target signal based on the cycle number, corrects the frequency according to a plurality of phase shift signals generated based on the reference signal, and minimizes an error of the calculated frequency by increasing the quantity of the phase shift signals, so as to enhance the accuracy of the calculated frequency of the target signal, speed up measurement, and reduce required circuit areas.

Abstract: A method for calibrating frequency, applicable to calibrating a frequency signal generated by a frequency generating unit of an apparatus at a preset frequency, includes obtaining the cycle number of the clock rate of a frequency signal based on a reference signal and a clock mask synchronous with the frequency signal; obtaining a frequency of the frequency signal based on the cycle number; correcting the frequency according to a plurality of phase shift signals generated based on the reference signal; and minimizing an error of the frequency of the frequency signal by increasing the quantity of the phase shift signals, so as to calibrate the frequency signal generated by the frequency generating unit.

Abstract: A method for measuring speed involves calculating and measuring speed of an object based on a distance and a time obtained by a method for measuring distance and a method for measuring time, respectively. The time between distance measuring sessions is obtained using the cycle number of a reference signal based on a clock mask synchronous with the distance measuring sessions. The time is corrected according to a plurality of phase shift signals generated based on the reference signal. An error of the time is minimized by increasing the quantity of the phase shift signals. The method enhances the accuracy of the measured time between distance measuring sessions, speeds up speed measurement, and reduces the required circuit areas. A system for measuring speed is further introduced for use with the method.