No abstract

No abstract

To explore the possibility of soft IP implementation, a fully digital smart temperature sensor without any full-custom device is proposed for painless VLSI or SOC on-chip integrations. The signal is processed thoroughly in time domain instead of conventional voltage or current domain. A cyclic delay line is used to generate the thermally sensitive pulse with a width proportional to the measured temperature. The timing reference is just the input clock, and a counter instead of voltage or current analog-to-digital converter is utilized for digital output coding. The circuit is realized by FPGA chips for functionality verification and performance evaluation. Implemented with as few as 140 Logic Elements, the proposed smart sensor was measured to have an error of -0.7°C40.9°C over a wide temperature range of -40°C-130°C. The effective resolution is better than 0.1°C, and the power consumption is 8.42pW at a sample rate of 2 samples/s. The performance is as good as those of most full-custom predecessors. The longest conversion time is around 260ps, and a conversion rate of 3 kHz at least is promised.Index Terms-temperature sensor, smart sensor, fully digital smart sensor, cyclic delay line, time domain, on-chip integration, low cost.

No abstract

In order to achieve the resolution comparable to that of a monolithic primary mirror telescope and make the imaging quality of the imaging system reach or approach the diffraction limit, the submirrors of the segments telescope should ensure co-phase splicing. To solve the problem of phase error detection, a high-precision piston error detection method based on convolutional neural network (CNN) is proposed. By setting a mask with a sparse multi-subpupil configuration on the exit pupil of the imaging system, a point spread function (PSF) image dataset that is extremely sensitive to the piston error is constructed. According to the characteristics of this dataset, a high-performance CNN model is built. And the best detection range of CNN is tested. The simulation results show that a single network can accurately output the piston error of one or more submirrors in the capture range slightly less than one wavelength. When applied to the six-submirror imaging system, the detection precision of the piston error reaches 0.0013λ RMS (Root Mean Square). And the method has good robustness to residual tip-tilt error, wavefront aberration, and CCD noise, light source bandwidth. The method is simple and fast, and can be widely used in the detection of the piston error of the segments.