Boot Block Encoding for NAND Data Retention at High Temperature
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Memory devices face data retention issues during the boot process, particularly in NAND memory devices, due to charge loss mechanisms, which can lead to common mode degradation and failure of boot information copies, especially at high temperatures, without advanced error handling capabilities.
Innovation Solution
Implementing a high reliability secondary encoding scheme for boot information, using multiple copies with bit inversion and additional spare bits, along with modified read levels and read retries, to enhance data reliability during the boot process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple copies of boot information are stored in NAND memory devices, then data redundancy is improved, but charge loss and common mode degradation still cause failure at high temperatures
Solution Approach 1:
The boot information is divided into multiple segments stored at different locations in the memory device. Each segment is independently encoded and can be recovered individually. The system segments the boot block into multiple pages distributed across different physical locations, preventing common mode degradation from affecting all copies simultaneously.
Solution Approach 2:
The patent applies different encoding parameters and read levels to different segments of boot information. By varying the encoding scheme and read parameters for each segment, the system compensates for charge loss variations at different memory locations and temperatures, ensuring at least one segment remains recoverable.
2Reliability
If advanced error handling capabilities are implemented, then data retention is improved, but device complexity increases
Solution Approach 1:
Error prevention measures are implemented during the write operation itself. The system encodes boot information with redundant parity bits and distributes segments across multiple locations before writing, rather than attempting complex error correction after retrieval fails. This preliminary encoding prevents errors rather than correcting them.
Solution Approach 2:
Instead of implementing complex error correction algorithms, the system creates multiple simplified copies of boot information with different encoding schemes. Each copy is stored independently with basic error detection codes, and the system simply reads through copies until one succeeds, avoiding complex real-time error handling.
3Productivity
If boot information is stored without encoding schemes, then write speed is improved, but data reliability deteriorates due to charge loss
Solution Approach 1:
Encoding is performed during the initial write operation rather than during read operations. The boot information is pre-encoded with redundancy bits and distributed across multiple locations in a single write pass, maintaining fast write speeds while ensuring future read reliability even under temperature stress.
Solution Approach 2:
The system uses different encoding parameters for different segments of boot information. Some segments use more robust encoding with higher redundancy, while others use lighter encoding. This varied parameter approach balances write overhead against read reliability requirements for different criticality levels.
Data Source
AI summary
A memory device includes a boot block that stores boot block code encoded using an encoding scheme. The boot block code includes a set of machine-readable instructions for booting the memory sub-system. A read command directed at the boot block is received while the memory device is in a boot state. Based on the command, the encoded boot block code is read from the boot block and decoded based on the encoding scheme. The decoded boot block code is provided to a memory sub-system controller responsive to the command