Memory Cell Self-Refresh Circuit for High-Energy Particle Upsets

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing memory devices in aerospace applications are prone to errors due to high-energy particle collisions, which can cause transient or permanent changes in memory cell states, and current solutions like triple modular redundancy are costly and complex to implement, especially when RF functions are involved.

Innovation Solution

A memory device using a D flip-flop with an RC retention circuit and Schmitt Trigger, along with an exclusive OR function and multiplexers, automatically corrects high-energy particle collisions without the need for a refreshing clock, allowing self-refreshing of data and reducing costs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If triple modular redundancy with voting element is used to correct high-energy particle collisions, then reliability is improved, but device complexity and cost increase

Engineering Contradiction:
Improveresistance to high-energy particle collisionsVSAvoidcomplexity of triple modular redundancy system
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent divides the memory cell into three separate functional copies (first, second, and third memory cells) that operate independently. Each copy processes data through its own path, allowing error detection and correction without requiring a complex centralized voting mechanism. This segmentation approach reduces overall system complexity while maintaining reliability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent creates three redundant copies of the memory cell circuitry and data paths. By copying the essential functional elements and comparing their outputs, the system can identify and correct single-event upsets without needing complex voting logic. The copying approach simplifies the architecture compared to traditional TMR voting systems.

Inventive Principle:
Principle #26Copying

2Reliability

If triple modular redundancy with voting element is used to correct high-energy particle collisions, then reliability is improved, but cost increases

Engineering Contradiction:
Improveresistance to high-energy particle collisionsVSAvoidmanufacturing cost of memory device
Core Design Contradiction:
ReliabilityVSEase of manufacture

Solution Approach 1:

By segmenting the memory function into three independent but simplified cell copies with direct comparison logic, the patent reduces the need for expensive voting circuitry. Each segment is simpler to manufacture, and the overall cost is reduced compared to traditional TMR implementations.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent uses functional copying of memory cells with direct output comparison, eliminating the need for complex voting elements. This copying approach with simplified error handling reduces manufacturing costs while maintaining the reliability benefits of redundancy.

Inventive Principle:
Principle #26Copying

3Reliability

If data refresh clock is used in triple modular redundancy system, then error correction is achieved, but RF functions are disrupted due to spectral pollution

Engineering Contradiction:
Improveerror correction capabilityVSAvoidspectral pollution from data refresh clock
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent extracts and eliminates the data refresh clock function from the system. Instead of using periodic clock signals to refresh and verify data, the system continuously monitors the three memory cell outputs and automatically detects and corrects errors without generating spectral pollution that would interfere with RF operations.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The memory system performs self-verification and self-correction by continuously comparing the outputs of three memory cells. This self-service mechanism eliminates the need for external refresh clocks, allowing the system to maintain reliability without generating harmful electromagnetic emissions that would disrupt RF functions.

Inventive Principle:
Principle #25Self-service

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 solution provides a cost-effective and efficient method to correct high-energy particle collisions in memory devices, ensuring robustness against such impacts without disrupting RF functions, and is easily implementable using both ASICs and discrete components.

Implementation Method 1

said retention means comprise an electronic circuit including an electrical resistor and a capacitor

Methodology Applied
Scientific EffectCapacitance: Capacitance

Implementation Method 2

An RC circuit for retaining the stored value in the memory cell is insensitive to high-energy particle collisions

Methodology Applied
Scientific EffectRC circuit retention:

Implementation Method 3

Using a Schmitt trigger allows the system output to transmit a ripple-free signal by eliminating internal voltage variations generated in the RC circuit due to high-energy particle collisions

Methodology Applied
Scientific EffectHysteresis: Hysteresis

Implementation Method 4

said detection means include means for performing an 'exclusive or' function

Methodology Applied
Scientific EffectLogical comparison:

Data Source

PatentEP2538414B1Memory device correcting the effect of high-energy particle collisions
Publication Date: 2019.08.28 THALES SA
  • EP2538414B1 patent drawingFigure 1~2
  • EP2538414B1 patent drawingFigure 3a~4
  • EP2538414B1 patent drawing

AI summary

The device has a retention unit (MRET) for retaining single copy of stored value (Qd) in a memory cell (CM) during specified time. A detection unit (MDET) detects a change of state of the memory cell by comparing the stored value in the memory cell with the retained value in the retention unit. A management unit (MG) determines whether a change of state of the memory cell is detected due to collision of a high energy particle. A loading of the value stored in the retention unit in the memory cell is automatically controlled when the change of state of the memory cell is detected. An independent claim is also included for a method for automatically correcting effect of collision of high-energy particles. (125).