搜索

x
中国物理学会期刊

双互锁存储单元触发器的置位冗余加固设计及单粒子翻转效应

CSTR: 32037.14.aps.75.20251574

Redundancy hardening design for single event upset tolerance in a dual interlocked storage cell flip-flop

CSTR: 32037.14.aps.75.20251574
PDF
HTML
导出引用
  • 标准单元库是芯片设计的基础, 对其进行有效的抗辐照加固设计与验证是保障宇航电子系统在轨可靠运行的关键. 本文以双互锁存储单元触发器为研究对象, 系统分析了其在置位冗余加固前后的单粒子翻转效应. 研究发现, 未加固的双互锁存储单元触发器在重离子入射时可能因置位信号受单粒子瞬态脉冲干扰而引发单粒子翻转; 而通过置位冗余加固后, 其两路置位信号可有效避免同时受到单粒子瞬态的干扰, 并利用双互锁存储单元结构的自恢复机制实现对节点电压异常的自动纠正, 显著提升了抗单粒子翻转能力. 本文冗余加固设计技术及单粒子仿真方法可为标准单元库的抗辐照设计及验证提供有效技术路径, 有助于缩短宇航元器件研发周期, 推动航空航天电子系统的发展.

     

    Ensuring on-orbit reliability of aerospace electronic systems requires effective radiation-hardened-by-design (RHBD) at the standard-cell level. This study addresses a less-studied single-event-upset (SEU) mechanism in a settable dual-interlocked storage-cell (DICE) flip-flop designed in SMIC 40 nm technology, which originates from its set-control path. A combined simulation approach integrates full-layout heavy-ion strike scanning using TREES with a 0.1 μm spatial grid and SPICE transient analysis driven by TCAD-calibrated single-event-transient (SET) current injection, thereby correlating layout-sensitive regions with circuit-critical nodes. Results indicate that heavy-ion strikes with a linear energy transfer (LET) of 37 MeV·cm2·mg–1 on the drain regions of set-circuit transistors P1 and N2 generate characteristic “101”-type SET pulses. Under the critical input condition where the data input D is at logic low and the set input SD is at logic high, these pulses transiently activate the normally-off set path, forcing the output Q from low to high and triggering SEU. To mitigate this vulnerability, we propose a set-redundancy RHBD scheme that utilizes two physically separated set-signal paths. The layout spacing minimizes the probability of concurrent transients, while the inherent self-recovery of the DICE core corrects single-node voltage deviations. Post-RHBD simulations show no SEU events under the tested conditions: the SEU cross-section drops from 1.7×10–9 cm2 in the baseline design, corresponding to 18 events out of 1691 strike locations, to zero in the RHBD version with 0 events out of 1919 strike locations. The introduced overhead remains moderate, with a 14.3% area increase and a 14.5% power increase, while critical timing parameters are largely unaffected. This work provides a concrete RHBD strategy and a correlated simulation framework for identifying and hardening latent weak points in radiation-hardened standard cells for aerospace integrated circuits.

     

    目录

    /

    返回文章
    返回