Memory ECC Using Analog Sensing and Configurable Error Probabilities
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Future high-density memory technologies, particularly Non-Volatile memory, face high raw bit error rates due to technology scaling, which existing error-correcting codes (ECC) struggle to address effectively, leading to design, development, and manufacturing challenges, as well as increased time to market and varying ECC requirements.
Innovation Solution
A customizable error correction system that includes an analog to digital sense unit, a configurable look-up table, and an error-correcting code (ECC) unit, which senses analog signals from memory cells, converts them into digital values, and updates error probability data based on operating conditions, allowing for flexible error correction and reduced Bit Error Rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional ECC machines are used for high density memory, then manufacturing cost is reduced, but error correction capability is insufficient for high raw bit error rates
Solution Approach 1:
The patent implements dynamic error probability updates through a look-up table that stores and updates error probabilities based on distributions of digital values from analog sensing. This allows the ECC system to adapt to changing memory characteristics over time and across different operating conditions, significantly improving error correction capability for high-density memory while maintaining cost-effectiveness through software-based updates rather than hardware redesigns
Solution Approach 2:
The system changes the parameter of error probability values stored in the look-up table based on observed error patterns and operating conditions. By dynamically adjusting these probability parameters, the ECC algorithm can optimize its correction performance for specific memory devices and conditions, resolving the contradiction between maintaining high reliability and avoiding expensive custom hardware designs
2Reliability
If different ECC requirements are implemented for different product developments, then error correction is optimized, but design and development time increases
Solution Approach 1:
The patent creates a universal ECC framework with a configurable look-up table that can serve multiple product developments and ECC requirements. The same base hardware platform can be used across different products by simply updating the error probability data in the look-up table, eliminating the need for separate design and development cycles for each product while maintaining optimized error correction for specific requirements
Solution Approach 2:
The system performs preliminary characterization of memory error patterns and pre-computes error probability values that are stored in the look-up table during manufacturing or initial operation. This preliminary action captures product-specific characteristics without requiring extensive post-manufacturing tuning, reducing design and development time for future product iterations
3Reliability
If analog sensing with soft-decision algorithms is used, then error correction performance is improved, but device complexity increases
Solution Approach 1:
The patent introduces a look-up table as an intermediary between the analog-to-digital sense unit and the ECC algorithm. This look-up table pre-stores error probability values that translate complex analog sensing data into simplified probability metrics, allowing the ECC algorithm to operate efficiently without directly processing complex analog signals, thus improving performance while managing device complexity
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system provides effective error correction by adapting to variations in voltage, temperature, and process conditions, significantly reducing Bit Error Rate and optimizing error correction for specific memory requirements, thereby enhancing the reliability and efficiency of high-density memory devices.
Implementation Method 1
an analog to digital sense unit which senses analog signals associated with the memory array cells and converts the analog signals into distributions of digital values
Data Source
AI summary
Described herein are a method and apparatuses for providing customizable error correction for memory arrays. In one embodiment, an apparatus includes a memory device having a memory array to store data and an analog to digital sense unit coupled to the memory array. The analog to digital sense unit senses analog signals associated with the memory array and converts the analog signals into distributions of digital values. An error-correcting code (ECC) unit receives the distributions of digital values from the analog to digital sense unit. A configurable non-volatile look-up table generates ECC parameters including error probability data and provides the ECC parameters to the ECC unit for error correction. The error probability data has error probability values that are associated with the distributions of digital values. The ECC unit executes an ECC algorithm to provide error correction using the error probability data.