Non-Volatile Memory Temperature-Aware Data Scrubbing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Non-volatile memory systems face increased bit error rates and decreased performance due to cross-temperature effects, where the threshold voltage of memory cells varies with temperature, leading to inefficient read voltages across different word lines in a block, especially when programmed at different temperatures.
Innovation Solution
An apparatus and method that records a temperature code in spare non-volatile memory cells on each word line, allowing for accurate readback even at extreme temperatures, and adjusts read voltage levels based on cross-temperature differences to reduce bit error rates and speed up the decoding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If memory cells are programmed at different temperatures without compensation, then programming flexibility is maintained, but bit error rate increases due to cross-temperature effects
Solution Approach 1:
The patent records the programming temperature in temperature tracking memory cells immediately after programming operations are completed on each word line. This preliminary recording of temperature data enables subsequent read operations to be compensated for cross-temperature effects, reducing bit error rates without adding complexity to the programming process itself.
Solution Approach 2:
The patent introduces temperature tracking memory cells as intermediary elements that store temperature information. These special memory cells act as mediators between the programming process and the read verification process, providing temperature compensation data that allows accurate reading of memory cells programmed at different temperatures.
2Measurement precision
If temperature compensation is implemented for all word lines, then read accuracy improves, but processing time increases due to additional verification steps
Solution Approach 1:
The patent performs temperature recording during the programming phase itself, so that when read operations occur later, the temperature compensation data is already available. This eliminates the need for time-consuming temperature measurements during read operations, maintaining fast decoding speeds while improving read accuracy.
Solution Approach 2:
The patent implements a feedback mechanism where the recorded temperature information from programming operations is used to adjust read voltage levels. This feedback loop enables the system to automatically compensate for temperature variations without manual intervention or additional processing time, improving both accuracy and efficiency.
3Reliability
If read voltage levels are optimized for each word line based on programming temperature, then data integrity improves, but system complexity increases due to multiple voltage levels
Solution Approach 1:
The patent applies different read voltage levels to different word lines based on their specific programming temperatures. Instead of using a single uniform read voltage for all word lines, the system tailors the read voltage to the local temperature conditions of each word line, improving data integrity while managing complexity through localized optimization.
Solution Approach 2:
The patent dynamically adjusts the read voltage parameter based on the programming temperature recorded for each word line. By changing the voltage parameter according to temperature conditions, the system optimizes data integrity without requiring fundamental architectural changes, managing complexity through parameter adaptation rather than structural complexity.
Data Source
AI summary
Techniques disclosed herein cope with cross-temperature effects in non-volatile memory systems. One technology disclosed herein includes an apparatus and method that scrubs a block of non-volatile memory cells responsive to a determination that variance in word line program temperatures in the block exceeds a threshold. Such blocks having large variance in programming temperatures for different word lines can potentially have high BERs when reading. This may be due to the difficulty in having one set of read levels that are optimum for all word lines in the block.