Fact-checked by Grok 2 weeks ago

Triple modular redundancy

Triple modular redundancy (TMR) is a fault-tolerance technique in digital systems design that replicates critical functional modules three times and uses majority voting on their outputs to detect and mask errors from a single faulty module, thereby maintaining system reliability in the presence of hardware failures. The foundational ideas for TMR emerged from John von Neumann's 1956 theoretical work on constructing reliable computing organisms from unreliable components, which introduced concepts of redundancy and error correction through multiplexing and voting mechanisms. This was followed by practical engineering analysis in a 1962 IBM study by N.H. Lyons and W. Vanderkulk, who demonstrated TMR's effectiveness against permanent component failures using Monte Carlo simulations on an IBM 704 computer, showing significant reliability improvements—for instance, achieving 95% reliability over 100 hours of operation with 60 modules each having a mean time to failure of 100 hours. TMR extends the broader N-modular redundancy (NMR) framework, where N=3 provides tolerance for up to one fault without system disruption, assuming independent module failures and a fault-free voter. In modern applications, TMR is extensively employed in safety-critical domains such as and to mitigate radiation-induced single-event upsets (SEUs) in field-programmable gate arrays (FPGAs) and other . implementations include variants like block TMR (BTMR), localized TMR (LTMR), and distributed TMR (DTMR), which triplicate logic at different granularities during or post-synthesis to ensure error masking while verifying functional equivalence. These approaches enable components to operate reliably in harsh environments, as seen in reprogrammable FPGAs like the ProASIC3E, though they introduce overheads in area, power, and performance—often exceeding 200% without optimization. Beyond , TMR enhances reliability in plants, medical devices, and automotive systems, prioritizing fault masking over in scenarios where is unacceptable.

Fundamentals

Definition and Purpose

Triple modular redundancy (TMR) is a specific instance of N-modular where N=3, employing three identical functional modules that execute the same task in parallel to enhance system reliability. This architecture replicates the core processing unit, such as processors or logic circuits, to provide against failures. The primary purpose of TMR is to achieve high reliability in systems by tolerating up to one faulty module without causing overall system failure, making it indispensable for safety-critical environments like space missions and medical devices. By duplicating modules, TMR ensures continued operation even if a single component experiences a fault, thereby minimizing the risk of catastrophic errors in applications where downtime or incorrect outputs could have severe consequences. At its core, TMR provides fault masking through a mechanism where a single fault in one module does not propagate to the system output, as long as the other two modules produce consistent results that can be selected via majority agreement. This conceptual approach isolates errors at the module level, preventing them from affecting the final system behavior and maintaining operational integrity. The reliability improvement offered by TMR can be quantified under assumptions of module failures and perfect , yielding the reliability R_{TMR} = 3p^2 - 2p^3, where p is the reliability of an individual . This demonstrates how TMR significantly boosts overall dependability compared to a single , particularly for modules with moderate reliability values.

Basic Principles

Triple modular redundancy (TMR) employs an consisting of three identical or diverse redundant modules that the same input independently and in parallel, with their outputs combined through a majority voting mechanism to produce a single system output. This duplication ensures that the system can continue operating correctly even if one module experiences a fault, as the voter selects the output shared by the majority of modules. The modules can be implemented in , such as circuits or processors, or in software, but the core principle relies on their independent execution to isolate potential errors. The fault model underlying TMR assumes that faults are independent and occur at most singly within the system during a given operational period, encompassing both transient faults (temporary errors due to noise or interference) and permanent faults (resulting from component degradation or failure). Under this model, TMR effectively masks a single fault by relying on the two unaffected modules to provide the correct output via majority vote, thereby maintaining system reliability without detecting or repairing the fault explicitly. However, TMR cannot mask simultaneous faults in two or more modules, as this would result in a tied or erroneous majority vote, potentially leading to system failure. The probability of TMR producing a correct output, assuming perfect and independent module failures, can be derived by considering the of correct and faulty modules. Let p denote the reliability of a single module (the probability it produces a correct output). The system succeeds if all three modules are correct, with probability p^3, or if exactly two are correct and one is faulty, with probability $3p^2(1-p), since the majority vote will select the correct output in the latter case. Thus, the overall system reliability R is given by: R = p^3 + 3p^2(1-p) = 3p^2 - 2p^3. This formula demonstrates TMR's fault coverage: for p > 0.5, R > p, meaning improves reliability, with maximum improvement near p \approx 0.5 where a single module is marginally reliable. For TMR to be effective, the modules must operate independently to minimize correlated errors, and while identical implementations are often used for simplicity and cost efficiency, diverse designs—such as different hardware or software variants—are recommended to avoid common-mode failures where a shared flaw affects all modules simultaneously. enhances by reducing the likelihood of identical faults propagating across modules, though it introduces challenges in and . This prerequisite ensures that the single-fault assumption holds in practice, preserving the derived reliability benefits.

Voting and Implementation

Majority Voting Logic

In triple modular redundancy (TMR), the majority voting logic employs a process that selects the output value agreed upon by at least two of the three redundant modules, thereby masking the effect of a single faulty module. For instance, if the modules produce outputs A, A, and B, the voter selects A as the system output, ensuring without requiring identification of the erroneous module. This rule leverages the to maintain correct operation as long as no more than one module fails. The core of this logic is the three-input majority function, a Boolean operation that outputs true if at least two inputs are true, and false otherwise. Mathematically, for inputs A, B, and C, it is expressed as: \text{MAJ}(A, B, C) = (A \land B) \lor (A \land C) \lor (B \land C) This function ensures an unambiguous result due to the binary nature of the outputs and the odd number of voters, preventing ties that could occur in even-numbered redundancy schemes. During TMR operation, the three modules compute the same function in on identical inputs, with their outputs simultaneously fed into the voter for resolution. This execution cycle allows fault masking, as the voter aggregates results without sequential dependencies between modules. The odd level inherently resolves decisions, supporting continuous system reliability in fault-prone environments. Beyond masking, the voting logic facilitates error detection by identifying discrepancies, such as when two agree but the third differs, which can trigger diagnostic signals or maintenance alerts. However, the primary function remains error concealment rather than correction, with detection serving as a secondary capability to indicate potential failures for further .

Voter Designs

The basic voter in triple modular redundancy (TMR) systems is a circuit that computes the majority output from three identical module inputs, typically implemented using AND and OR gates to form the expression \text{OUT} = (A \land B) \lor (A \land C) \lor (B \land C), where A, B, and C are the inputs. This design ensures that the output matches the majority value among the three inputs, masking single faults in one module. The truth table for this three-input majority voter is as follows:
Inputs (A, B, C)Output
0, 0, 00
0, 0, 10
0, 1, 00
0, 1, 11
1, 0, 00
1, 0, 11
1, 1, 01
1, 1, 11
This table confirms the majority function across all combinations, with the circuit requiring only four gates (three AND and one OR) in a standard implementation. Advanced voter designs incorporate fault detection to enhance reliability beyond basic masking, such as self-checking voters that use dual voter instances and XOR gates to signal mismatches between their outputs, thereby detecting voter faults themselves. For instance, one self-checking approach employs two parallel majority voters followed by a with XOR logic to generate an error indicator if discrepancies occur, often implemented with clocked flip-flops for in synchronous systems. Asynchronous voters, in contrast, avoid clocks to reduce timing hazards and power in certain applications, using quasi-delay-insensitive logic to compute early outputs while maintaining hazard-free operation. To prevent the voter from becoming a , it must itself be implemented with , such as triplicating the voter circuit and applying majority logic across the triplicate outputs, which limits error propagation and ensures system-level fault masking even if one voter fails. This approach has been integrated into integrated circuits () for radiation-hardened applications, where voters are replicated at the gate level during to distribute . In hardware implementations, voters introduce minimal , typically around 0.1 ns due to two logic levels in basic designs or single complex gates in optimized variants, while power overhead remains low at 1-3 μW per voter under standard-cell synthesis. Advanced designs like those using OR-AND complex gates can reduce area to about 3.3 μm² and power to 1.1 μW, offering better figures of merit for overall TMR efficiency compared to traditional AND-OR configurations.

Historical Development

Origins in Early Computing

The conceptual foundations of triple modular redundancy (TMR) emerged in the mid-20th century amid efforts to build reliable computing systems from inherently unreliable components, particularly during the 1950s and 1960s when fault-tolerant computing theory was formalized. John von Neumann's seminal work, presented in lectures in 1952 and published posthumously in 1956 as Probabilistic Logics and the Synthesis of Reliable Organisms Operating Essentially As Clocks from Unreliable Components, explored the use of redundancy and multiplexing to mitigate errors in automata constructed from probabilistic elements. This theory demonstrated that by replicating logical units and employing voting mechanisms, systems could achieve high reliability despite component failure rates as high as 1 in 100 operations, influencing subsequent designs in fault-tolerant architectures. NASA played a pivotal role in advancing TMR through its adoption in the Apollo program, where reliability was paramount for manned spaceflight. In the 1960s, TMR was integrated into the Saturn V rocket's Instrument Unit (IU), a critical guidance and control system developed by IBM under NASA contract. The IU's Launch Vehicle Digital Computer (LVDC) employed TMR across its seven logic stages, triplicating modules and using majority voting to mask faults, thereby achieving a mission reliability of approximately 0.996 for a 250-hour flight profile—significantly higher than the 0.955 of non-redundant designs. This implementation targeted transient and permanent errors in the guidance processors, including those induced by cosmic rays in the space environment, ensuring robust performance during launch and orbital insertion. Early adoption of TMR faced substantial challenges due to the hardware constraints of the era, particularly the high cost and complexity of triplicating circuits in systems reliant on vacuum tubes and early transistors. In the vacuum-tube period of the , full TMR demanded excessive resources, often leading to hybrid approaches that combined duplication for less critical paths with TMR for vital functions to balance reliability and expense. These trade-offs highlighted the technique's promise but underscored the need for advancing component densities and integration before widespread feasibility.

Key Milestones

A pivotal early milestone was the 1962 study by N.H. Lyons and W. Vanderkulk at , which provided the first practical engineering analysis of TMR. Using simulations on an computer, they demonstrated TMR's effectiveness against permanent component failures, showing significant reliability improvements—for instance, achieving 95% reliability over 100 hours of operation with 60 modules each having a mean time to failure of 100 hours. The successful launch of in 1969, which enabled the first human , relied in part on triple modular redundancy (TMR) implemented in the rocket's Launch Vehicle Digital Computer (LVDC) for fault-tolerant guidance and control during ascent. This application of TMR, featuring triplicated logic stages with majority voting at each pipeline phase, masked potential single-event faults and contributed to the mission's reliability without hardware failures. In the and , TMR saw integration into military systems to enhance in high-stakes environments, including assessments for advanced VLSI designs that employed TMR alongside two-rail networks for fault masking in electronics. Concurrently, advanced fault-tolerant mainframes to achieve high reliability for scientific and business computing. These developments established TMR as a standard for mitigating transient faults in operational systems. The 1990s marked TMR's standardization in (FPGA) designs, particularly through (now ) tools for radiation-hardened space applications, where TMR mitigated single-event upsets (SEUs) in configurable logic. A seminal from the analyzed reconfigurable TMR systems with spares, differentiating permanent and intermittent faults to improve reliability models for dynamic environments. From the 2000s onward, TMR became widespread in and processor ecosystems, especially for SEU mitigation in FPGA-integrated designs, using triplicated hardware instances with to prevent in safety-critical applications. A 2023 study in Nature Scientific Reports examined TMR reliability under step-stress accelerated life tests, highlighting parameter estimation for systems with partial fault coverage using the . In 2024, research proposed an enhanced TMR (ETMR) model for fault-tolerant systems, integrating additional redundancy layers to boost uptime in urban infrastructure. The 2010s also witnessed a shift toward software-based TMR in , where triple replication of application instances with majority ensured in distributed environments like data centers.

Applications

Aerospace and Avionics

Triple modular redundancy (TMR) plays a critical role in and systems, where high reliability is essential to mitigate risks from environmental factors such as cosmic radiation and mechanical failures during flight. In space applications, TMR is widely employed in satellites and deep-space probes to protect processors and circuits against single-event upsets caused by cosmic rays, which can induce transient errors in . For instance, the Perseverance Mars rover utilizes TMR within its field-programmable gate arrays (FPGAs), where three identical circuit copies perform computations in parallel, and a voter selects the majority output to mask radiation-induced faults, ensuring uninterrupted operation in the harsh Martian radiation environment. In , TMR enhances in systems certified by the (FAA) for commercial aircraft, ensuring precise control of flight surfaces even under fault conditions. The employs triple-redundant flight control computers that process pilot inputs via majority voting, maintaining stability and preventing catastrophic failures in critical maneuvers. This architecture aligns with FAA requirements for redundancy in electronic flight control systems, where TMR helps achieve failure rates below 10^{-9} per flight hour for primary controls. NASA's onboard computers also leveraged TMR principles across three redundant channels in the primary system, allowing the vehicle to continue operations if one channel failed, as demonstrated during multiple missions. LayerZero Power Systems integrates TMR into uninterruptible power supplies () for launch vehicles and platforms, providing three independent control paths that deliver up to two orders of magnitude improvement in system reliability compared to single-module designs, safeguarding power distribution during high-vibration launches. Despite these benefits, TMR implementation in hardware introduces challenges, including increased weight and power consumption due to the replication of components, which is particularly constraining in weight-sensitive flight systems. To address this, engineers often apply partial TMR, focusing only on critical paths such as radiation-vulnerable processors or actuators, thereby balancing reliability with resource efficiency. For example, in designs, selective TMR on FPGAs reduces overhead while still mitigating effects effectively. These trade-offs highlight TMR's adaptability in mission-critical environments, where partial implementations can achieve substantial without excessive penalties.

Computing and FPGAs

Triple modular redundancy (TMR) has been implemented at the processor level to enhance reliability in environments susceptible to soft errors, such as systems. In the soft from /, the TMR subsystem provides comprehensive soft error detection, correction, and recovery by triplicating the processor cores and integrating dedicated TMR managers, voters, and comparators. This approach ensures in radiation-prone applications, allowing seamless operation without system downtime upon error detection. In field-programmable gate arrays (FPGAs), TMR is widely applied to configurable logic blocks (CLBs) to create radiation-tolerant designs, where logic functions are triplicated and outputs are voted to mask single event upsets (SEUs). This technique mitigates configuration memory errors in SRAM-based FPGAs, commonly used in space and high-radiation settings, by combining TMR with periodic scrubbing for sustained reliability. Intel's FPGA solutions, such as those in the Agilex and families, incorporate TMR as a core SEU mitigation strategy to prevent bit flips in logic and memory elements without interrupting system function. Hierarchical TMR (H-TMR) extends basic TMR in FPGAs by applying redundancy at multiple levels, forming networks that interconnect triplicated modules for improved (IC) fault tolerance. This method enhances error masking through layered voting, reducing the probability of system failure in complex designs. A 2009 probabilistic model for H-TMR networks demonstrates that fault coverage improves significantly with hierarchical depth, achieving up to 99.999% reliability in modeled ICs under random fault assumptions. Software-based TMR supports high-availability computing clusters by replicating computations across multiple , akin to mechanisms for data but applied to processing tasks for fault detection via majority voting. This approach enables error correction in distributed systems without hardware modifications, particularly useful for mission-critical software in unreliable environments. Notable examples include the Saturn launch vehicle's central in its Instrument Unit, which employed TMR across seven logic stages with majority voting to ensure guidance reliability during ascent. In modern contexts, TMR is integrated into accelerators for deep space missions, such as 's SpaceCube Low Power Edge Resilient , which uses tripled Edge TPUs to tolerate radiation-induced faults in onboard inference. Recent implementations as of 2024 include TMR hardening of softcores for embedded aerospace applications and triplex redundant flight control systems based on the M1394B bus.

Advantages, Limitations, and Variants

Reliability Benefits and Trade-offs

Triple modular redundancy (TMR) provides high fault coverage by masking single failures through majority voting, achieving system reliability of approximately 99.999% when individual reliability is 0.999, based on the R_{\text{TMR}} = 3R^2 - 2R^3. This masking enables continuous operation in safety-critical applications without immediate downtime, as the system remains functional as long as at least two modules agree. For instance, TMR can extend (MTBF) by one to two orders of magnitude compared to non-redundant systems, potentially shifting reliability from hours of operation to years in demanding environments like . Despite these gains, TMR incurs significant trade-offs, including a threefold increase in hardware cost and power consumption due to the replication of modules and addition of voting logic. It also introduces latency from the voter, which can impact performance in time-sensitive systems. Furthermore, TMR is vulnerable to common-mode faults—shared failures across modules due to identical designs or environmental factors—if diversification measures like varied implementations are not applied, potentially reducing effective fault tolerance. In comparison to a single module, which offers no fault tolerance and fails immediately upon any error, TMR provides robust masking of single faults. Dual modular redundancy (DMR), while requiring only 2x resources, detects but does not automatically mask single faults, often necessitating manual intervention or system halt, whereas TMR ensures seamless correction. A 2023 study in analyzed TMR under partially accelerated life tests, modeling repairable faults and highlighting degradation in reliability estimation if the voter experiences failures, as it assumes perfect voter operation for optimal masking; real-world voter faults can thus compromise overall system resilience.

Extensions like Hierarchical TMR

Hierarchical triple modular redundancy (H-TMR) extends basic TMR by applying recursively across modules and voters, enabling in large-scale digital integrated circuits (ICs). In H-TMR, TMR is implemented not only at the bit or gate level but also at higher structural levels, such as grouping modules into triplicated hierarchies where voters themselves are protected by additional TMR layers. This recursive approach masks faults more effectively in complex systems by propagating majority voting through multiple tiers, reducing the probability of system-wide failure due to cascading errors. A probabilistic model for analyzing H-TMR's fault masking performance demonstrates that it achieves substantially higher reliability than conventional TMR, particularly for high-volume digital systems where fault propagation risks increase with scale. Enhanced TMR (ETMR) builds on TMR by incorporating spare modules and dynamic reconfiguration for self-healing capabilities, allowing the system to replace faulty components without full shutdown. In ETMR architectures, redundant spares are activated upon voter detection of discrepancies, combined with Markov processes to model and predict state transitions for zero- operation. This variant has been applied to fault-tolerant systems, where ETMR ensures continuous operation by automatically reconfiguring around hardware failures, achieving near-elimination of operational downtime compared to standard TMR setups. Other extensions include N-modular redundancy schemes where N exceeds 3, such as quadruple modular redundancy (QMR), which uses four identical modules and a voter to tolerate up to two simultaneous faults, offering greater resilience in environments with higher error rates than basic TMR can handle. QMR reduces overhead relative to full replication while improving fault coverage, often through approximate integrations that balance reliability and efficiency in resource-constrained designs. Recent advancements as of 2025 include efficient TMR techniques for enhancing reliability in deep neural networks (DNNs) against bit-flip faults using explainable , and low-power hardened latches based on TMR for triple-node-upset recovery in radiation environments. These extensions collectively enhance TMR's scalability and fault coverage; for instance, H-TMR's hierarchical structure increases tolerance to multiple concurrent faults in FPGA-based systems by mitigating voter vulnerabilities that limit basic TMR, resulting in markedly improved overall system reliability for radiation-prone applications.

References

  1. [1]
    [PDF] The Use of Triple-Modular Redundancy to Improve Computer ...
    This paper is concerned with system failures caused by permanent component fail- ures, in contrast to the problem of transient failures caused by noise, which ...Missing: origin | Show results with:origin
  2. [2]
    Concepts of Self-Repairing Systems - Sage Journals
    Jul 19, 2013 · 4. Von Neumann J. Probabilistic logics and the synthesis of reliable organisms from unreliable components. Automata Studies 1956; 34: 43–98.
  3. [3]
    [PDF] have introduced the concept of N-modular Redundancy (NMR ...
    TMR – Triple Modular Redundancy. - most well known version of NMR – Figure 4.1. - typically assume that all modules are identical in terms of digital logic.
  4. [4]
    [PDF] Verification of Triple Modular Redundancy Insertion for Reliable and ...
    Binary decision diagram (BDD). •. Block Triple Modular Redundancy (BTMR). •. Combinatorial logic (CL). •. Computer Aided Design (CAD).
  5. [5]
    [PDF] Localized Triple Modular Redundancy vs. Distributed Triple Modular ...
    Field programmable gate arrays (FPGA) are used in every space application. Currently, most space flight applications use- radiation hardened (RH) FPGAs, ...
  6. [6]
    [PDF] Energy-Efficient Permanent Fault Tolerance in Hard Real-Time ...
    Triple Modular Redundancy (TMR) is an NMR system through which three copies of tasks2 are executed, and the result that is produced by two or all of the ...Missing: definition | Show results with:definition
  7. [7]
    Reliability analysis of the triple modular redundancy system under ...
    Sep 7, 2023 · Triple modular redundancy (TMR) is a robust technique utilized in safety-critical applications to enhance fault-tolerance and reliability.Missing: history | Show results with:history
  8. [8]
    Triple modular redundancy | ACM SIGDA Newsletter
    Triple modular redundancy (TMR) is a fault tolerant form of N-modular redundancy, in which three systems perform a process and that result is processed by a ...
  9. [9]
    [PDF] A Primer on Architectural Level Fault Tolerance
    Feb 1, 2008 · Triple Modular Redundancy (TMR): A fault-tolerant architecture that uses three processors and voting to produce the output. Voting: the ...
  10. [10]
    [PDF] SmallSat Avionics - NASA
    Triple Modular Redundancy (TMR) is a fault mitigation technique where ... It is used in space flight, medical, networking and many more embedded devices.
  11. [11]
    [PDF] ECE 313 — Probability with Engineering Applications Fall 2000 ...
    Triple Modular Redundancy (TMR). ○ Triple modular redundancy (TMR) systems ... (3p2 – 2p3) ≈ 3×10–8 and the probability of error is ≈ q. ECE 313 ...
  12. [12]
  13. [13]
    [PDF] On the Optimal Design of Triple Modular Redundancy Logic ... - arXiv
    The TMR scheme uses three identical logic blocks performing the same task in tandem with corresponding outputs being compared through majority voters.
  14. [14]
    [PDF] A Fault Tolerance Improved Majority Voter for TMR System ... - arXiv
    TMR, which forms a subset of N-modular redundancy, requires two-times duplication of a function module and the three identical function modules are joined ...
  15. [15]
    1.5. Triple Modular Redundancy - Intel
    If an SEU affects one of the hardware instances, the voting logic notes the majority output. This operation masks malfunctioning hardware. With TMR, your design ...
  16. [16]
    [PDF] Verification of Triple Modular Redundancy (TMR) Insertion for ...
    Fault tolerance (FT). •. Field programmable gate array (FPGA). •. Formal verification (FV). •. Global triple modular redundancy (GTMR). •. Hardware description ...
  17. [17]
    FPGA-based design of a self-checking TMR voter - IEEE Xplore
    Abstract: The most common error mitigation scheme used for hardening designs against radiation-induced upsets on FPGAs is Triple Modular Redundancy (TMR).
  18. [18]
    [PDF] Area Optimized Quasi Delay Insensitive Majority Voter for TMR ...
    Recently, quasi delay insensitive (QDI) asynchronous designs of a 3-input majority voter based on different QDI logic synthesis methods were explored and ...
  19. [19]
    [PDF] Redundancy in fault tolerant computing
    Voter: if the Voter fails, the complete system fails. Voter is a single point of failure. Triplicated Voters in a TMR configuration. Page 7. 7. The effect of ...
  20. [20]
    [PDF] Power, Delay and Area Comparisons of Majority Voters relevant to ...
    In this paper, a number of standard-cell based majority voter designs relevant to TMR architectures are presented, and their power, delay and area parameters ...
  21. [21]
    [PDF] Lectures on PROBABILISTIC LOGICS AND THE SYNTHESIS OF ...
    The following lists the versions of the von Neumann lectures that exist at Caltech or that have been published. There are 2 versions at Caltech and 3 ...
  22. [22]
    [PDF] Fault-Tolerant Computing: A Preamble for Assuring Viability of Large ...
    fault-tolerant computing and the value of component redundancy was delivered by John von. Neumann during five lectures (ref. 2). In the 1950's redundancy was ...
  23. [23]
    [PDF] Description and performance of the saturn launch vehicle's ...
    TRIPLE MODULAR REDUNDANCY. General description of the data adapter. The data adapter is the input/output unit for the digital computer and communicates with ...
  24. [24]
    [PDF] Radiation Tolerant Computing for Aerospace Applications
    Jul 20, 2009 · Cosmic radiation induces transients in integrated circuits. - Commercial processes are susceptible to Singe-Event-Upsets (SEUs).Missing: ray | Show results with:ray<|control11|><|separator|>
  25. [25]
    Design of fault-tolerant computers - ACM Digital Library
    J Von Neumann Probabilistic logics and the synthesis of reliable organisms ... Basic concepts, motivation, and techniques of fault tolerance are discussed in this ...
  26. [26]
    [PDF] Application of Redundancy in the Saturn V Guidance and Control ...
    The Instrument Unit of the Saturn Vvehicle is basically equivalent to that ... A triplex, or triple modular redundant (TMR), arrangement is shown in Figure 3.
  27. [27]
    [PDF] Fault Tolerant VLSI Design Assessments for Advanced Avionics ...
    Triple modular redundancy (TM) and two rail networks. (TPR\) are techniques used in fault masking. These designs have a high redundant ratio i.e. ratio of ...
  28. [28]
    [PDF] Fault Tolerant Hardware - Duke University
    – IBM mainframes. – IBM xServers. – Sun UltraEnterprise (E6000, E10000, E12000 ... – TMR, pair-and-spare, etc.: all use lots of hardware and power.
  29. [29]
    [PDF] AEROSPACE AND DEFENSE - AMD
    For traditional TMR, this means that state machines have to be reset to fully recover from a SEU. Xilinx TMR solves this problem by inserting voters on all ...Missing: processors | Show results with:processors
  30. [30]
    Reliability analysis of triple modular redundancy system with spare
    Dec 1, 1993 · In this work, a reconfigurable Triple Modular Redundancy (TMR) with spare system that differentiates between permanent and intermittent faults ...
  31. [31]
    Reliability analysis of the triple modular redundancy system under ...
    Sep 7, 2023 · This article focuses on estimating the distribution parameters of a TMR system under step-stress partially accelerated life tests.
  32. [32]
    [PDF] Application of Enhanced Triple Modular Redundancy (ETMR ...
    Oct 23, 2024 · Hence this study developed an Enhanced Fault-tolerant Traffic Light System (EFTLS) Model that utilizes an Enhanced Triple Modular Redundancy. ( ...
  33. [33]
    Radiation resistance is baked into the Perseverance Mars rover ...
    May 6, 2021 · On Perseverance, the circuits are "triple-module redundant," which means that Xilinx manufactures three copies of each circuit powering an ...Missing: cosmic | Show results with:cosmic
  34. [34]
    Apollo Guidance Computer and the First Silicon Chips
    Oct 14, 2015 · The Apollo Guidance Computer used early silicon chips, tested rigorously, and had no hardware failures during missions. The chips were tested ...
  35. [35]
    Why Airbus Has Such A Significant Safety Track Record With Fly-By ...
    Sep 18, 2025 · ... Airbus system with side-stick controls and triple-redundant computers. Today, fly-by-wire is the industry standard and has coincided with a ...
  36. [36]
    Triple Modular Redundancy - LayerZero Power Systems
    Triple Modular Redundancy is the most popular static redundant architecture for large, complex systems. • TMR scales perfectly and is completely general in ...
  37. [37]
    MicroBlaze Triple Modular Redundancy (TMR) Subsystem (PG268)
    Describes the Triple Modular Redundancy IP Cores. The complete MicroBlaze™ TMR solution, providing soft error detection, correction and recovery for Xilinx ...Missing: SEU | Show results with:SEU
  38. [38]
    Overview - 1.0 English - PG268
    The Xilinx ® Triple Modular Redundancy (TMR) solution for the Vivado® Design Suite is designed for these applications, providing all the necessary building ...
  39. [39]
    Design Automation for Radiation Tolerant FPGAs | Synopsys Blog
    Sep 27, 2022 · Some of the automated mechanisms are triple modular redundancy for memory, I/O and logic, safe state machines, and memory error correction ...
  40. [40]
    [PDF] Hierarchical Triple-Modular Redundancy (H-TMR) Network ... - arXiv
    Abstract: Hierarchical application of Triple-Modular Redundancy (TMR) increases fault tolerance of digital Integrated Circuit (IC). In this paper, a simple ...
  41. [41]
    (PDF) Hierarchical Triple-Modular Redundancy (H-TMR) Network ...
    Hierarchical application of Triple-Modular Redundancy (TMR) increases fault tolerance of digital Integrated Circuit (IC). In this paper, a simple probabilistic ...
  42. [42]
    Software-Based Triple Modular Redundancy with Fault-Tolerant ...
    Our Scalable Software Support for Dependable Embedded Systems (S3DES) approach achieves fault tolerance by utilizing software-based triple modular redundancy ...Missing: clusters | Show results with:clusters
  43. [43]
    [PDF] Current AI Technology in Space
    AI in space is used for low-latency predictions, autonomous applications, remote sensing, data triage, and autonomous rover controls.
  44. [44]
    1.5. Triple Modular Redundancy - Intel
    Triple modular redundancy (TMR) is an established SEU mitigation technique for improving hardware fault tolerance. ... For example, Synopsys* offers a tool that ...
  45. [45]
    Area, Power and Speed Optimized Early Output Majority Voter for ...
    Jun 14, 2021 · This paper presents a new, efficient asynchronous early output majority voter that can be used to effectively realize an asynchronous triple modular redundancy ...
  46. [46]
    Reliability modeling of TMR computer systems with repair and ...
    The effect of repair and common-mode failures on the reliability and mean up time of a TMR computer configuration is analyzed and shown to be significant.
  47. [47]
    Hierarchical Triple-Modular Redundancy (H-TMR) Network For ...
    Feb 2, 2009 · In this paper, a simple probabilistic model was proposed for analysis of fault masking performance of hierarchical TMR networks.
  48. [48]
    Application of Enhanced Triple Modular Redundancy (ETMR ...
    Oct 14, 2024 · This study developed an Enhanced Fault-tolerant Traffic Light System (EFTLS) Model that utilizes an Enhanced Triple Modular Redundancy (ETMR) architecture and ...
  49. [49]