Periodic Receiver Training for Memory I/O Timing Drift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Unmatched memory device architectures improve bandwidth and frequency but suffer from degraded timing control due to non-self-compensating delays, which are exacerbated by environmental changes and aging, leading to potential link failures and performance degradation.
Innovation Solution
Implementing periodic training to detect errors in a training signal and adjust I/O parameters, allowing for continuous compensation of timing drift, thereby maintaining the sampling strobe signal within the optimal 'eye' of the data signal, even under temperature and voltage variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If unmatched architecture is used to improve bandwidth and frequency, then receiver power and performance are improved, but timing control is degraded due to non-self-compensating delays
Solution Approach 1:
The patent implements a feedback mechanism where the receiver detects timing errors in received data and communicates these errors back to the transmitter. The transmitter then adjusts its timing parameters based on this feedback to compensate for drift, resolving the timing control degradation issue while maintaining the power benefits of unmatched architecture
Solution Approach 2:
The system performs self-adjustment of timing parameters through automated error detection and correction. The receiver monitors its own timing performance and triggers retraining sequences when drift is detected, allowing the system to self-correct without external intervention
2Manufacturing precision
If periodic training is implemented to compensate for timing drift, then timing control is improved, but bus bandwidth and performance are impacted
Solution Approach 1:
Instead of continuous training, the patent implements periodic training sequences that are triggered only when timing drift exceeds a threshold. This periodic approach maintains timing accuracy while minimizing the impact on bus bandwidth compared to continuous training methods
Solution Approach 2:
The system performs partial retraining only when necessary, based on detected timing errors. Rather than continuously retraining or training all parameters, the system selectively retrains only when timing drift is detected, reducing the overall training overhead and preserving bus bandwidth
3Use of energy by moving object
If tight timing parameters are used to reduce power, then power consumption is reduced, but I/O latency is more negatively affected by temperature and voltage variation
Solution Approach 1:
The patent implements dynamic timing parameters that can be adjusted based on operating conditions. The system starts with tight timing for power efficiency and dynamically relaxes or adjusts parameters when drift is detected, maintaining both power efficiency and reliability under varying temperature and voltage conditions
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
I/O parameters are adjusted based on a number of errors detected in a received training signal. A controller device sends the training signal while a memory device is in a training mode. The memory device samples the training signal and the system causes an adjustment to at least one I/O parameter based on a detected number of errors. Either the controller or the memory device can perform the error detection, depending on the configuration of the system. Either an I/O parameter of the controller or an I/O parameter of the memory device can be adjusted, depending on the configuration of the system.