Analyzing the influence on Q factor, which was caused by the parasitic effect in a CMOS RF on chip integrated inductor, a concise method to increase the Q factor has been obtained when optimizing the layout parameter....Analyzing the influence on Q factor, which was caused by the parasitic effect in a CMOS RF on chip integrated inductor, a concise method to increase the Q factor has been obtained when optimizing the layout parameter. Using this method, the Q factor of 7.9 can be achieved in a 5nH inductor (operating frequency is 2GHz) while the errors in inductance are less than 0.5% compared with the aimed values. It is proved by experiments that this method can guarantee the sufficient accuracy but require less computation time. Therefore, it is of great use for the design of the inductor in CMOS RF IC’s.展开更多
提高射频功率器件的鲁棒性有利于增强器件的抗静电放电能力和抗失配能力.为了直观地了解器件内部发生的电学过程,本文研究了高鲁棒性N型沟道RF-LDMOS (Radio Frequency Lateral Diffusion MOS)在TLP (Transmission Line Pulse)应力下的...提高射频功率器件的鲁棒性有利于增强器件的抗静电放电能力和抗失配能力.为了直观地了解器件内部发生的电学过程,本文研究了高鲁棒性N型沟道RF-LDMOS (Radio Frequency Lateral Diffusion MOS)在TLP (Transmission Line Pulse)应力下的电学机理.利用0.18μm BCD (Bipolar/CMOS/DMOS)先进制程,实现了特定尺寸器件的设计与流片.通过实测与仿真的对比,发现静电放电失效的随机性、芯片内部的热效应是导致仿真和实测差异的非理想因素.通过对TLP仿真的各阶段重要节点的分析,证明了源极下方的P型埋层有利于提高空穴电流的泄放能力,从而提高RF-LDMOS的鲁棒性.展开更多
随着信号采集设备的带宽越来越宽,大量感兴趣或者不感兴趣信号被捕捉,多信号的盲识别问题是一个难题,更是一个亟需解决的问题。传统的识别大都基于功率、频谱或相位等诸多先验知识进行模板匹配,但在全盲条件下对多信号进行自适应识别是...随着信号采集设备的带宽越来越宽,大量感兴趣或者不感兴趣信号被捕捉,多信号的盲识别问题是一个难题,更是一个亟需解决的问题。传统的识别大都基于功率、频谱或相位等诸多先验知识进行模板匹配,但在全盲条件下对多信号进行自适应识别是一个更加复杂的问题。为此,提出了一种基于通用软件无线电外设(Universal Software Radio Peripheral,USRP)、片上射频网络(RF Network on Chips,RFNOC)和Keras的自适应信号盲识别算法。首先构造基于深度学习的神经网络,然后使用初始IQ数据、初始功率谱密度数据和快速傅里叶变换(Fast Fourier Transform,FFT)累积算法处理后的谱相关密度数据等三种不同的初始数据去训练它,利用其自适应性实现多信号的盲识别,最后通过基于USRP、RFNOC和Keras的软硬件验证了该算法的有效性和鲁棒性。展开更多
文摘Analyzing the influence on Q factor, which was caused by the parasitic effect in a CMOS RF on chip integrated inductor, a concise method to increase the Q factor has been obtained when optimizing the layout parameter. Using this method, the Q factor of 7.9 can be achieved in a 5nH inductor (operating frequency is 2GHz) while the errors in inductance are less than 0.5% compared with the aimed values. It is proved by experiments that this method can guarantee the sufficient accuracy but require less computation time. Therefore, it is of great use for the design of the inductor in CMOS RF IC’s.
文摘随着信号采集设备的带宽越来越宽,大量感兴趣或者不感兴趣信号被捕捉,多信号的盲识别问题是一个难题,更是一个亟需解决的问题。传统的识别大都基于功率、频谱或相位等诸多先验知识进行模板匹配,但在全盲条件下对多信号进行自适应识别是一个更加复杂的问题。为此,提出了一种基于通用软件无线电外设(Universal Software Radio Peripheral,USRP)、片上射频网络(RF Network on Chips,RFNOC)和Keras的自适应信号盲识别算法。首先构造基于深度学习的神经网络,然后使用初始IQ数据、初始功率谱密度数据和快速傅里叶变换(Fast Fourier Transform,FFT)累积算法处理后的谱相关密度数据等三种不同的初始数据去训练它,利用其自适应性实现多信号的盲识别,最后通过基于USRP、RFNOC和Keras的软硬件验证了该算法的有效性和鲁棒性。
