|
|
Frontiers of Information Technology & Electronic Engineering
ISSN 2095-9184 (print), ISSN 2095-9230 (online)
2024 Vol.25 No.2 P.316-322
A low-noise, high-gain, and large-dynamic-range photodetector based on a JFET and a charge amplifier
Abstract: We demonstrate a low-noise, high-gain, and large-dynamic-range photodetector (PD) based on a junction field-effect transistor (JFET) and a charge amplifier for the measurement of quantum noise in Bell-state detection (BSD). Particular photodiode junction capacitance allows the silicon N-channel JFET 2sk152 to be matched to the noise requirement for charge amplifier A250. The electronic noise of the PD is effectively suppressed and the signal-to-noise ratio (SNR) is up to 15 dB at the analysis frequency of 2.75 MHz for a coherent laser power of 50.08 μW. By combining of the inductor and capacitance, the alternating current (AC) and direct current (DC) branches of the PD can operate linearly in a dynamic range from 25.06 μW to 17.50 mW. The PD can completely meet the requirements of SNR and dynamic range for BSD in quantum optics experiments.
Key words: Quantum noise; Bell-state detection (BSD); Photodetector (PD); Junction field-effect transistor (JFET); Charge amplifier
机构：浙江工业大学，机械工程学院，中国杭州，310023
目的：针对现有多保真建模方法需要嵌套训练样本来捕捉数据相关性导致的计算成本增加与未充分考虑不同保真度样本之间非线性关系导致模型精度低的问题，本文提出一种结合多保真建模和主动学习的可靠性分析方法，旨在实现高效且准确的失效概率估计。
创新点：1.基于非线性自回归方案构建了一种非线性自回归多保真克里金（NAMK）模型；2.在模型更新过程中，用集成的多保真学习函数代替传统的学习函数，通过综合考虑采样成本和多保真样本之间的相关性，从多保真样本空间中选择新的采样点；3.当选择高保真样本时，使用残差模型生成嵌套的低保真样本。
方法：1.在指定的参数范围内选择初始多保真样本，并使用NAMK构建初始代理模型；2.通过集成学习函数确定新样本的位置和保真度；3.一旦选择了一个高保真样本，根据残差模型生成嵌套的低保真度样本并根据新的样本更新模型；4.使用基于相对误差估计的停止准则终止主动学习过程并输出失效概率估计结果。
结论：1.本文所提出的基于多保真建模和主动学习的可靠分析方法提高了失效概率估计的效率和精度；2.利用NAMK模型来捕捉多保真样本之间的非线性关系，有效提高了代理模型的准确性；3.考虑多保真样本的相关性和采样成本的学习函数能自适应地确定新样本的位置和保真度；4.当学习函数选择高保真样本时，通过构造残差模型生成嵌套的低保真度样本可减少低保真度模型的调用次数。
关键词组：
References:
Open peer comments: Debate/Discuss/Question/Opinion
<1>
DOI:
10.1631/FITEE.2300340
CLC number:
TN29
Download Full Text:
Downloaded:
1358
Download summary:
<Click Here>Downloaded:
392Clicked:
2228
Cited:
0
On-line Access:
2024-08-27
Received:
2023-10-17
Revision Accepted:
2024-05-08
Crosschecked:
2023-10-08