Refine
Has Fulltext
- no (19)
Language
- English (19)
Is part of the Bibliography
- yes (19) (remove)
Keywords
- DMR (2)
- Fault tolerance (2)
- TMR (2)
- radhard design (2)
- reliability (2)
- (SET) count rate (1)
- ASIC (1)
- Absorbed dose (1)
- Analytical models (1)
- Answer set programming (1)
The relentless improvement of silicon photonics is making optical interconnects and networks appealing for use in miniaturized systems, where electrical interconnects cannot keep up with the growing levels of core integration due to bandwidth density and power efficiency limitations. At the same time, solutions such as 3D stacking or 2.5D integration open the door to a fully dedicated process optimization for the photonic die. However, an architecture-level integration challenge arises between the electronic network and the optical one in such tightly-integrated parallel systems. It consists of adapting signaling rates, matching the different levels of communication parallelism, handling cross-domain flow control, addressing re-synchronization concerns, and avoiding protocol-dependent deadlock. The associated energy and performance overhead may offset the inherent benefits of the emerging technology itself. This paper explores a hybrid CMOS-ECL bridge architecture between 3D-stacked technology-heterogeneous networks-on-chip (NoCs). The different ways of overcoming the serialization challenge (i.e., through an improvement of the signaling rate and/or through space-/wavelength division multiplexing options) give rise to a configuration space that the paper explores, in search for the most energy-efficient configuration for high-performance.
A hybrid design approach of the hierarchical physical implementation design flow is presented and demonstrated on a fault-tolerant low-power multiprocessor system. The proposed flow allows to implement selected submodules in parallel with contrary requirements such as identical placement and individual block implementation. The overall system contains four Leon2 cores and communicates via the Waterbear framework and supports Adaptive Voltage Scaling (AVS) functionality. Three of the processor core variants are derived from the first baseline reference core but implemented individually at block level based on their clock tree specification. The chip is prepared for space applications and designed with triple modular redundancy (TMR) for control parts. The low-power performance is enabled by contemporary power and clock management control. An ASIC is fabricated in a low-power 0.13 mu m BiCMOS technology process node.
Use of a standard non-rad-hard digital cell library in the rad-hard design can be a cost-effective solution for space applications. In this paper we demonstrate how a standard non-rad-hard flip-flop, as one of the most vulnerable digital cells, can be converted into a rad-hard flip-flop without modifying its internal structure. We present five variants of a Triple Modular Redundancy (TMR) flip-flop: baseline TMR flip-flop, latch-based TMR flip-flop, True-Single Phase Clock (TSPC) TMR flip-flop, scannable TMR flip-flop and self-correcting TMR flipflop. For all variants, the multi-bit upsets have been addressed by applying special placement constraints, while the Single Event Transient (SET) mitigation was achieved through the usage of customized SET filters and selection of optimal inverter sizes for the clock and reset trees. The proposed flip-flop variants feature differing performance, thus enabling to choose the optimal solution for every sensitive node in the circuit, according to the predefined design constraints. Several flip-flop designs have been validated on IHP's 130nm BiCMOS process, by irradiation of custom-designed shift registers. It has been shown that the proposed TMR flip-flops are robust to soft errors with a threshold Linear Energy Transfer (LET) from (32.4 MeV.cm(2)/mg) to (62.5 MeV.cm(2)/mg), depending on the variant.
The radiation-sensitive field-effect transistors (RADFETs) with an oxide thickness of 400 nm are irradiated with gate voltages of 2, 4 and 6 V, and without gate voltage.
A detailed analysis of the mechanisms responsible for the creation of traps during irradiation is performed.
The creation of the traps in the oxide, near and at the silicon/silicon-dioxide (Si/SiO2) interface during irradiation is modelled very well. This modelling can also be used for other MOS transistors containing SiO2.
The behaviour of radiation traps during postirradiation annealing is analysed, and the corresponding functions for their modelling are obtained. The switching traps (STs) do not have significant influence on threshold voltage shift, and two radiation-induced trap types fit the fixed traps (FTs) very well. The fading does not depend on the positive gate voltage applied during irradiation, but it is twice lower in case there is no gate voltage.
A new dosimetric parameter, called the Golden Ratio (GR), is proposed, which represents the ratio between the threshold voltage shift after irradiation and fading after spontaneous annealing. This parameter can be useful for comparing MOS dosimeters.
The soft error rate (SER) due to heavy-ion irradiation of a clock tree is investigated in this paper. A method for clock tree SER prediction is developed, which employs a dedicated soft error analysis tool to characterize the single-event transient (SET) sensitivities of clock inverters and other commercial tools to calculate the SER through fault-injection simulations. A test circuit including a flip-flop chain and clock tree in a 65 nm CMOS technology is developed through the automatic ASIC design flow. This circuit is analyzed with the developed method to calculate its clock tree SER. In addition, this circuit is implemented in a 65 nm test chip and irradiated by heavy ions to measure its SER resulting from the SETs in the clock tree. The experimental and calculation results of this case study present good correlation, which verifies the effectiveness of the developed method.
A triple modular redundancy (TMR) based design technique for double cell upsets (DCUs) mitigation is investigated in this paper. This technique adds three extra self-voter circuits into a traditional TMR structure to enable the enhanced error correction capability. Fault-injection simulations show that the soft error rate (SER) of the proposed technique is lower than 3% of that of TMR. The implementation of this proposed technique is compatible with the automatic digital design flow, and its applicability and performance are evaluated on an FIFO circuit.
In this paper, an asynchronous design for soft error detection and correction in combinational and sequential circuits is presented. The proposed architecture is called Asynchronous Full Error Detection and Correction (AFEDC). A custom design flow with integrated commercial EDA tools generates the AFEDC using the asynchronous bundled-data design style. The AFEDC relies on an Error Detection Circuit (EDC) for protecting the combinational logic and fault-tolerant latches for protecting the sequential logic. The EDC can be implemented using different detection methods. For this work, two boundary variants are considered, the Full Duplication with Comparison (FDC) and the Partial Duplication with Parity Prediction (PDPP). The AFEDC architecture can handle single events and timing faults of arbitrarily long duration as well as the synchronous FEDC, but additionally can address known metastability issues of the FEDC and other similar synchronous architectures and provide a more practical solution for handling the error recovery process. Two case studies are developed, a carry look-ahead adder and a pipelined non-restoring array divider. Results show that the AFEDC provides equivalent fault coverage when compared to the FEDC while reducing area, ranging from 9.6% to 17.6%, and increasing energy efficiency, which can be up to 6.5%.
This paper presents a scalable E-band radar platform based on single-channel fully integrated transceivers (TRX) manufactured using 130-nm silicon-germanium (SiGe) BiCMOS technology. The TRX is suitable for flexible radar systems exploiting massive multiple-input-multipleoutput (MIMO) techniques for multidimensional sensing. A fully integrated fractional-N phase-locked loop (PLL) comprising a 39.5-GHz voltage-controlled oscillator is used to generate wideband frequency-modulated continuous-wave (FMCW) chirp for E-band radar front ends. The TRX is equipped with a vector modulator (VM) for high-speed carrier modulation and beam-forming techniques. A single TRX achieves 19.2-dBm maximum output power and 27.5-dB total conversion gain with input-referred 1-dB compression point of -10 dBm. It consumes 220 mA from 3.3-V supply and occupies 3.96 mm(2) silicon area. A two-channel radar platform based on full-custom TRXs and PLL was fabricated to demonstrate high-precision and high-resolution FMCW sensing. The radar enables up to 10-GHz frequency ramp generation in 74-84-GHz range, which results in 1.5-cm spatial resolution. Due to high output power, thus high signal-to-noise ratio (SNR), a ranging precision of 7.5 mu m for a target at 2 m was achieved. The proposed architecture supports scalable multichannel applications for automotive FMCW using a single local oscillator (LO).
In this paper two new methods for the design of fault-tolerant pipelined sequential and combinational circuits, called Error Detection and Partial Error Correction (EDPEC) and Full Error Detection and Correction (FEDC), are described. The proposed methods are based on an Error Detection Logic (EDC) in the combinational circuit part combined with fault tolerant memory elements implemented using fault tolerant master-slave flip-flops. If a transient error, due to a transient fault in the combinational circuit part is detected by the EDC, the error signal controls the latching stage of the flip-flops such that the previous correct state of the register stage is retained until the transient error disappears. The system can continue to work in its previous correct state and no additional recovery procedure (with typically reduced clock frequency) is necessary. The target applications are dataflow processing blocks, for which software-based recovery methods cannot be easily applied. The presented architectures address both single events as well as timing faults of arbitrarily long duration. An example of this architecture is developed and described, based on the carry look-ahead adder. The timing conditions are carefully investigated and simulated up to the layout level. The enhancement of the baseline architecture is demonstrated with respect to the achieved fault tolerance for the single event and timing faults. It is observed that the number of uncorrected single events is reduced by the EDPEC architecture by 2.36 times compared with previous solution. The FEDC architecture further reduces the number of uncorrected events to zero and outperforms the Triple Modular Redundancy (TMR) with respect to correction of timing faults. The power overhead of both new architectures is about 26-28% lower than the TMR. (C) 2015 Elsevier Ltd. All rights reserved.
In this paper two new methods for the design of fault-tolerant pipelined sequential and combinational circuits, called Error Detection and Partial Error Correction (EDPEC) and Full Error Detection and Correction (FEDC), are described. The proposed methods are based on an Error Detection Logic (EDC) in the combinational circuit part combined with fault tolerant memory elements implemented using fault tolerant master–slave flip-flops. If a transient error, due to a transient fault in the combinational circuit part is detected by the EDC, the error signal controls the latching stage of the flip-flops such that the previous correct state of the register stage is retained until the transient error disappears. The system can continue to work in its previous correct state and no additional recovery procedure (with typically reduced clock frequency) is necessary. The target applications are dataflow processing blocks, for which software-based recovery methods cannot be easily applied. The presented architectures address both single events as well as timing faults of arbitrarily long duration. An example of this architecture is developed and described, based on the carry look-ahead adder. The timing conditions are carefully investigated and simulated up to the layout level. The enhancement of the baseline architecture is demonstrated with respect to the achieved fault tolerance for the single event and timing faults. It is observed that the number of uncorrected single events is reduced by the EDPEC architecture by 2.36 times compared with previous solution. The FEDC architecture further reduces the number of uncorrected events to zero and outperforms the Triple Modular Redundancy (TMR) with respect to correction of timing faults. The power overhead of both new architectures is about 26–28% lower than the TMR.