摘要：This paper proposes a novel robust design method for the sense mode of a MEMS vibratory gyroscope based on fuzzy reliability and Taguchi design. The principles of fuzzy reliability and Taguchi design are both introduced and described in detail. Experimental results demonstrate that the signal to noise ratio of the robust design scheme is better than those of the other experimental *** the full temperature range from -40 to 80°C, the temperature sensitivities of phase margin, gain margin, sensitivity margin,the maximum amplitude of open loop system, bandwidth of closed loop system, and the performance function of the robust design system are all smaller than those of the original design system. Meanwhile, the temperature sensitivity of the bandwidth of the robust design system is improved to 126 from 1075 ppm/°C. Moreover, the bias drift over the full temperature range of the robust design system is improved to 61°/h from 179°/h.

摘要：Experimental evidence is presented showing obvious azimuthal dependence of single event upsets(SEU) and multiple-bit upset(MBU) patterns in radiation hardened by design(RHBD) and MBU-sensitive static random access memories(SRAMs), due to the anisotropic device layouts. Depending on the test devices, a discrepancy from 24.5% to 50% in the SEU cross sections of dual interlock cell(DICE) SRAMs is shown between two perpendicular ion azimuths under the same tilt angle. Significant angular dependence of the SEU data in this kind of design is also observed, which does not fit the inverse-cosine law in the effective LET method. Ion trajectory-oriented MBU patterns are identified, which is also affected by the topological distribution of sensitive volumes. Due to that the sensitive volumes are periodically isolated by the BL/BLB contacts along the Y-axis direction, double-bit upsets along the X-axis become the predominant configuration under normal *** triple-bit upset and quadruple-bit upset patterns are the same under different ion azimuths(Lshaped and square-shaped configurations, respectively). Those results suggest that traditional RPP/IRPP model should be promoted to consider the azimuthal and angular dependence of single event effects in certain *** earth-based evaluation of SEE sensitivity, worst case beam direction, i.e., the worst case response, should be revealed to avoid underestimation of the on-orbit error rate.