Archives

  • 2018-07
  • 2018-10
  • 2018-11
  • 2019-04
  • 2019-05
  • 2019-06
  • 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2019-12
  • br Introduction The endeavor to achieve

    2019-11-30


    Introduction The endeavor to achieve a successful use of commercial-off-the-shelf (COTS) electronic devices in space is not new and the reasons are well known [1], [2], [3]. More recently, COTS have been widely employed in the manufacturing of low-cost, small satellites, for scientific and even commercial applications, in Low-Earth-Orbit (LEO), short to medium-term space missions [4]. The use of COTS on board satellites requires, mandatorily, the addressing of degrading mth1 pathway effects like Total Ionizing Dose (TID) and Single-Event Latchup (SEL) [5], [6], [7]. The usual mitigation approaches are, respectively, metallic shielding and latchup current limiters. The addition of soft-error mitigation is also important since a single bit-flip in a register can result in a high-severity failure. Even though their higher LET (Linear Energy Transfer) threshold and lower saturation cross-section (ratio of upsets to the particle fluence) due to the hardening processes, rad-hard (radiation hardened) devices can also be, in a smaller scale, affected by soft errors. Consequently, the need of mitigation shall be evaluated in a case-to-case basis. Fault injection and susceptibility analysis tools can be used to assess the efficacy of the soft-error mitigation strategy. Simulation-based methods are well known to be time-consuming options and there are different ways to accelerate test execution [8], [9], [10], [11], however the scope and the extra effort and cost in programming or hardware development shall be taken into account. The study carried on in Ref. [12] is concerned with verifying the correct implementation of TMR to protect a generic netlist against SEUs (Single-Event Upset) and there is no evaluation about susceptibility improvement. In Ref. [13] Hamming was employed to achieve an effective protection of a generic FIR filter against SEUs. Susceptibility improvement evaluations were performed through the Single Event Upset Simulation Tool (SST), which is based on elaborated scripts of simulation tool commands to address the injection. The characterization and the modeling of SETs (Single-Event Transient) is an issue [14] that is not considered in related works. Besides studying a solution for an onboard tolerant system, this work evaluates the effectiveness of the soft-error mitigation strategy through simulation-based analysis of the VHDL (Very high-speed integrated circuits Hardware Description Language) code. The system employs the JPEG-LS image compression standard on account of the good tradeoff between compression rate, image loss and area requirements attained with a SRAM-based FPGA [15]. The moderate area and memory requirements made it possible to consider future use of two FPGA types: a flash ProASIC3 or an antifuse Axcelerator. These are medium-size devices with a lower susceptibility to the radiation effects than that of SRAM-based FPGAs [16], [17], and with enough room to house the hole compression system. The low memory demanded by predictive-differential compression methods, like JPEG-LS, is one important advantage over transform-based ones [15], but one drawback is that a single soft error can corrupt part of a packed data block [18]. Therefore, the addition of soft-error mitigation is mandatory and widespread TMR (Triple Modular Redundancy) and Hamming EDAC (Error Detection And Correction) were selected as fault mitigation techniques, taking into account an LEO space radiation environment. An upgraded version of the simulation-based method, described in Ref. [19], was employed to analyze the soft-error susceptibility of the JPEG-LS VHDL code. The susceptibility analysis method is based on random SET and SEU injection, and resulting error rate evaluation. In Ref. [19], a simplified JPEG-LS description (Regular mode only, without encoding process), handling a small test image (200 pixels per line), was submitted to a partial hardening analysis to verify the coherence of the proposed method and estimates.