A fully integrated step-up dc-dc converter for energy harvesting applications is presented. A minimum start-up voltage of 100 mV is achieved by the start-up mechanism based on an LC oscillator (LCO). Conventional voltage converters with the LCO-based start-up mechanism provide a significantly low conversion efficiency of around 1%. To overcome this drawback, the following two circuit techniques are proposed in this brief: 1) a gate-boosted charge pump; and 2) an LCO deactivation. The proposed voltage converter is fabricated in the 65-nm silicon-on-thin-buried-oxide process. The measurement results show that when the input voltage is 100 mV, the proposed converter can provide an output of 760 mV at a conversion efficiency of 33%, which is the highest efficiency compared with the conventional fully integrated voltage converters.
26 October 2016 in IEEE Transactions on Circuits and Systems II: Express Briefs
This brief presents an ultra-low power wake-up receiver using ultrasound for IoT applications. To achieve both high sensitivity and low power consumption, we propose a Colpitts-oscillator-based super-regenerative receiver (COSR). Owing to the simple architecture of the proposed COSR, a lowest supply voltage operation of 0.3 V and a smallest area are achieved. Furthermore, the power consumption of the proposed wake-up receiver is scalable and is determined by the input signal sensitivity and data rate, which are configurable by the user. In a field test, the proposed wake-up receiver consumes 1 $\mu$ W, which is smaller by 77% than that of a conventional ultrasound wake-up receiver at a comparable communication distance and data rate.
This study investigates the origin of low-frequency (LF) 1/ f noise in Si n-channel metal-oxide-semiconductor field-effect transistors (n-MOSFETs) under cryogenic operation. The fluctuation of the drain current increased with decreasing temperature, exhibiting LF 1/ f noise of more than two orders of magnitude higher at 2.5 K compared with that at 300 K. As revealed by the temperature dependence of the normalized current spectral density, the LF 1/ f noise at 2.5 K is primarily governed by carrier number fluctuations. To obtain insight into the carrier trapping centers under cryogenic operation, we investigate the effect of oxide/Si interface states on the LF 1/ f noise by utilizing Si n-MOSFETs with different surface orientations, i.e., different interface trap densities (D it ). The LF 1/ f noise is comparable between the surface orientations at 300 K, whereas excess noise was observed at 2.5 K for the surface orientation with higher D it in the order of (100)<(120)≤(110)-orientations. This indicates that the LF 1/ f noise at cryogenic temperatures originates from oxide/Si interface defects and disorders, that is, the interface states and band tail states. These states are localized at the conduction-band edge, which contributes to noise generation as the Fermi level approaches the conduction-band edge at cryogenic temperatures. This study demonstrates the significance of the oxide/Si interface quality in suppressing the LF 1/ f noise in Si MOS devices operated at cryogenic temperatures.
27 March 2017 in Japanese Journal of Applied Physics
We have demonstrated the operation of a CMOS inverter consisting of Si tunnel FinFETs. Both p- and n-type tunnel FinFETs are successfully fabricated and operated on the same SOI wafer. The current mismatch between p- and n-type tunnel FETs is compensated by tuning the number of fin channels. Very low short-circuit current and clear voltage input–output characteristics are obtained. The thin epitaxial channel in the tunnel FinFETs effectively increases the drain current and accordingly reduces the drain capacitance, which could help high-performance tunnel FET CMOS inverter operation.
29 September 2023 in IEEE Transactions on Circuits and Systems I: Regular Papers
In this paper, we propose a CMOS current comparison circuit operating at cryogenic temperature to readout spin quantum bit (qubit) states via a charge sensing scheme. A current integrator and correlation double sampling circuit are implemented in the proposed current comparator since precise current sensing is required to achieve readout the states of qubits. The measurement results indicate the feasibility that the reduced readout time by a factor of 1/100 and 100-times better current resolution with the readout fidelity of 99.9 % compared to the conventional readout systems are achieved. Additionally, the current and noise specifications for a charge sensor required to attain it are discussed. Furthermore, the proposed current comparator attains the best tradeoff among power, current resolution, and readout time, compared with the conventional cryogenic current readout circuits.
In this paper, a table lookup-based computing technique is proposed to perform convolutional neural network (CNN) inference without multiplication, and its FPGA implementation is demonstrated as a proof-of-concept. Conventionally, the hardware specific to the lookup-based dot product approximation (LDA) has been proposed to achieve energy-efficient CNN computations for edge AI applications. However, it has not been applied to complicated AI tasks, such as ImageNet image classification, because LDA degrades the inference accuracy, especially in complicated tasks. Therefore, a new LDA technique using residual vector quantization, called RLDA, is proposed in this study. By adopting the proposed RLDA to CNN, we can achieve an inference accuracy degradation of below 5% for ResNet-18 model on ImageNet 1000 classification. In addition, the proposed RLDA-based CNN accelerator is implemented on a ZCU104 evaluation board, which includes a Zynq UltraScale+ FPGA and DDR4 DRAM. We compare the processing time of the proposed accelerator on ZCU104 with that of NVIDIA Jetson AGX Orin and reveal that the processing time of the proposed accelerator is comparable to that of Jetson Orin in the lower layers of the ResNet-18 model. Finally, the computational performance of the proposed accelerator is compared with conventional FPGA-based accelerators. The proposed accelerator does not require a digital signal processor (DSP), which is often available in modern FPGAs. We demonstrate that the proposed accelerator achieves a performance more than four times higher than that of the conventional FPGA-based accelerator without DSP.