Dynamic Latch Leakage Compensation for Longer Data Retention

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Dynamic latches face challenges in retaining data due to charge leakage, leading to the need for frequent refresh cycles, which increases power consumption and may not be suitable for all devices, especially those requiring high-speed operations and low power consumption.

Innovation Solution

A dynamic latch design with a minimal number of transistors that includes compensation circuits to extend data retention time, reducing the need for frequent refresh cycles and lowering power consumption by maintaining data validity for a longer interval.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If dynamic latches are used to achieve high-speed operations and low power consumption, then operating speed and power efficiency are improved, but charge leakage occurs leading to short data retention time requiring frequent refresh cycles

Engineering Contradiction:
Improveoperating speedVSAvoiddata retention time
Core Design Contradiction:
SpeedVSDuration of action of stationary object

Solution Approach 1:

The patent applies preliminary anti-action by introducing compensation transistors (first and second compensation transistors) that proactively counteract charge leakage effects before data becomes invalid. These transistors are configured to compensate for charge loss in the memory node, extending the retention time of stored data without requiring frequent refresh cycles, thus resolving the contradiction between high-speed operation and short data retention time

Inventive Principle:
Principle #9Preliminary anti-action

2Reliability

If refresh rate is increased to maintain data validity before leakage causes data to become invalid, then data retention reliability is improved, but power consumption increases

Engineering Contradiction:
Improvedata validityVSAvoidpower consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent implements self-service through compensation circuits that automatically detect and correct charge leakage in the memory node without external intervention. The compensation transistors are controlled by control signals that activate them based on the actual charge state, allowing the latch to self-correct and maintain data validity continuously, thereby reducing the need for frequent high-power refresh cycles while maintaining reliability

Inventive Principle:
Principle #25Self-service

3Duration of action of stationary object

If static latches are used to avoid refresh cycles and extend data retention time, then data retention time is improved, but device size and power consumption increase

Engineering Contradiction:
Improvedata retention timeVSAvoidlatch size
Core Design Contradiction:
Duration of action of stationary objectVSDevice complexity

Solution Approach 1:

The patent merges the advantages of dynamic latches (small size, high speed, low power) with the advantage of static latches (long data retention) by combining dynamic latch structure with compensation circuits. The compensation transistors are integrated into the dynamic latch architecture, creating a hybrid structure that achieves extended retention time without the large size penalty of traditional static latches

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS20250006234A1Leakage compensated dynamic latch
Publication Date: 2025.01.02 INTEL CORP
  • US20250006234A1 patent drawing
  • US20250006234A1 patent drawing
  • US20250006234A1 patent drawing

AI summary

Some embodiments include input stage of a latch to receive input data information and clock information; a memory node coupled to the input stage to store information based on the input data information; an output stage of the latch coupled to the memory node and including an output node to provide output data information based on the information stored at the memory node; a first circuit to provide a first circuit path between the memory node and a first node in the input stage; and a second circuit to provide a second circuit path between the memory node and a second node in the input stage.