Memory DCA Training Using Multi-Phase Clock Eye Window Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor memory devices face challenges in achieving high signal integrity and efficient duty cycle adjustment due to time delays in setting optimal duty cycles during DCA training, and potential failures when command even gaps are not observed.
Innovation Solution
A training method that adjusts the eye window of data signals using a duty cycle adjuster (DCA) code by selecting specific DCA codes corresponding to internal clock signals with phase differences of 180°, 90°, and 270° relative to a reference internal clock signal, allowing for accurate measurement and adjustment of eye window sizes in units of one or two unit intervals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional DCA training is used to adjust duty cycle, then signal integrity can be improved, but training time increases excessively
Solution Approach 1:
The patent segments the DCA training process into two distinct phases: a first training operation that measures eye window size in units of two unit intervals, and a second training operation that measures eye window size in units of one unit interval. This segmentation allows the system to perform coarse adjustment quickly first, then fine-tune the duty cycle, thereby reducing overall training time while maintaining signal integrity.
Solution Approach 2:
The first training operation performs preliminary duty cycle adjustment by measuring eye window size in larger units (two unit intervals), establishing a baseline duty cycle setting before the more time-consuming second training operation refines the adjustment in smaller units (one unit interval). This preliminary action reduces the search space for the second training, decreasing total training time.
2Productivity
If DCA code is applied without observing command even gap, then training speed increases, but application may fail
Solution Approach 1:
The patent performs preliminary checks and preparations before applying DCA codes, ensuring that command even gaps are properly observed. The training sequence is structured to verify timing conditions beforehand, preventing failed applications while maintaining efficient training speed through the optimized two-stage measurement approach.
3Measurement precision
If eye window measurement is performed in fine units, then measurement precision increases, but measurement time increases
Solution Approach 1:
The measurement process is segmented into two stages: first measuring eye window size in units of two unit intervals for quick coarse measurement, then measuring in units of one unit interval for precise fine-tuning. This segmentation achieves high measurement precision while minimizing total measurement time by avoiding fine-grained measurement throughout the entire training process.
Solution Approach 2:
The patent applies partial fine-grained measurement only when necessary - specifically in the second training operation after the first operation has established a baseline. This partial application of high-precision measurement reduces overall measurement time while maintaining sufficient precision for duty cycle optimization.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A training method of a memory device adjusting an eye window of a data signal in response to a duty cycle adjuster (DCA) includes performing a first training operation (S120) that selects a first DCA code corresponding to a first internal clock signal (IBCLK) having a phase difference of 180° relative to a reference internal clock signal, and performing a second training operation (S130) that selects a second DCA code and a third DCA code respectively corresponding to a second internal clock signal (QCLK) and a third internal clock signal (QBCLK) having a phase difference of 90° and 270° relative to reference internal clock signal. In the first training operation (S120), the eye window size of the data signal is measured in units of two unit intervals, and in the second training operation (S130), the eye window size of the data signal is measured in units of one unit interval.