A 400 $\times$ 200 600fps 117.7dB-DR SPAD X-Ray Detector with Seamless Global Shutter and Time-Encoded Extrapolation Counter.
ISSCC(2023)
摘要
X-ray detectors using hard X-rays
$(\lambda < 0.1\text{nm})$
are widely used in industrial and medical equipment, and their market size has increased in recent years [1]–[3]. There is a great demand for real-time X-ray detectors to deliver high-frame-rate, and high-dynamic-range (HDR) images. However, conventional approaches such as switching modes between frames or combining images from different exposures, make X-ray images susceptible to motion artifacts, and require multiple exposures with X-ray sources. Therefore, to overcome these limitations, a global-shutter X-ray detector is required [4], which also should operate seamlessly due to the slow response of X-ray sources. To meet these requirements of X-ray detectors while providing high-quality images with low radiation dose, a photon-counting detector using single-photon avalanche diodes (SPADs) is now seen as a promising candidate [5], [6]. Generally, the X-ray detectors do not use lenses, and hence the size of the detector must be as large as their targets, which poses additional challenges. Therefore, no clinical-grade SPAD X-ray detector has been reported yet. This paper presents a photon-counting SPAD X-ray detector with seamless global shutter. To minimize the power consumption of the SPAD pixels, a time-encoded extrapolation counter is proposed, which minimizes the number of control signals and recycles the counter cells when they overflow, achieving 18b outputs only using 10b hardware. Two compact counters are implemented in a pixel to provide a seamless global shutter. The detector, implemented in 65nm CMOS, has a pixel pitch of
$49.5\mu \mathrm{m}$
and a pixel resolution of
$400\times 200$
, resulting in a chip size of
$24\times 14\text{mm}^{2}$
. It achieves a frame rate of 600fps and a dynamic range of 117.7dB, while consuming 127.2mW.
更多查看译文
AI 理解论文
溯源树
样例
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要