Memory Device Reference Voltage Generation for DRAM Signal Integrity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor memory devices, particularly DRAM, face reliability issues due to variations in physical characteristics of data lines, leading to signal variations and increased bit errors during data exchange.
Innovation Solution
A memory device method and system that generates reference voltage sets for each data line based on specific codes, including information about reference voltages and decision feedback equalization levels, to accurately sample and determine input data, thereby reducing bit errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single reference voltage is used for all data lines, then device complexity is reduced, but measurement precision deteriorates due to signal variations on different data lines
Solution Approach 1:
The patent segments the reference voltage generation by creating separate reference voltage sets for each data line (first reference voltage set for first data line, second reference voltage set for second data line). This segmentation allows each data line to have its own optimized reference voltage, improving measurement precision while managing complexity through systematic organization.
Solution Approach 2:
The patent applies local quality by generating reference voltages tailored to specific data lines based on their individual characteristics. Each data line receives a reference voltage set optimized for its physical properties, ensuring local optimization of signal integrity and reducing bit errors on each specific line.
2Measurement precision
If reference voltage sets are generated for each data line, then data sampling accuracy is improved, but device complexity increases
Solution Approach 1:
The patent implements dynamic reference voltage generation where the reference voltage sets are adjusted based on decision feedback equalization levels specific to each data line. This dynamic adaptation allows the system to optimize sampling accuracy for each line's characteristics while using a unified generation mechanism that manages complexity.
Solution Approach 2:
The patent changes the parameters of reference voltages (voltage levels, timing) based on data line characteristics and DFE levels. By systematically varying these parameters for each data line, the patent achieves high sampling accuracy without requiring fundamentally different circuit architectures, thus managing complexity.
3Adaptability or versatility
If data lines have different physical characteristics, then manufacturing flexibility is improved, but reliability deteriorates due to signal variations
Solution Approach 1:
The patent addresses reliability issues by applying local quality optimization - each data line receives a reference voltage set specifically tuned to its physical characteristics. This local optimization compensates for variations in data line properties, ensuring reliable data transmission across all lines despite their different characteristics.
Solution Approach 2:
The patent uses decision feedback equalization levels as a feedback mechanism to adjust reference voltage sets for each data line. This feedback loop allows the system to adapt to actual signal conditions on each line, improving reliability by compensating for physical characteristic variations through systematic adjustment.
Data Source
AI summary
A memory device may include a first data line driver circuit that generates a first reference voltage set based on a first code and a second code associated with a first data line, and determines bit values of the first input data received through the first data line, based on the first reference voltage set. A second data line driver circuit may similarly generate a second reference voltage set. The reference voltages may have levels based on a decision feedback equalization (DFE) technique to reduce bit errors otherwise caused by inter symbol interference.