GaN HEMT微波器件大信号统计模型研究
本文选题:GaN + HEMT ; 参考:《电子科技大学》2016年博士论文
【摘要】:随着无线通信、雷达和电子战等领域的快速发展,人们对微波收发机的性能提出了更高的要求。功率放大器作为发射机的关键组件,其中最主要、最关键的是功率晶体管,其带宽、输出功率、效率和工作温度等性能将严重影响系统的整体性能。与以硅(Si)为代表的第一代半导体材料和以砷化镓(GaAs)为代表的第二代半导体材料相比,第三代半导体材料氮化镓(GaN)具有禁带宽度大、饱和电子速度高、临界击穿电场高和热导率高等优势,所以它已成为半导体器件的一个重要的研究课题。GaN高电子迁移率晶体管(HEMT)的模型的准确性对电路设计的影响非常大。目前,针对GaN HEMT的物理模型和大、小信号等效电路模型都已取得一定进展,然而由于生产过程中的工艺误差、模型的不确定性、工艺参数的变化以及环境因素的影响,器件及电路的一致性和工艺分析与优化设计显得格外重要,因此开展GaN微波器件统计模型研究具有重要意义。本文针对GaN HEMT等效电路模型展开研究,建立了大、小信号等效电路统计模型;并以等效电路统计模型为基础,设计了GaN HEMT高效率功率放大器对所建立的统计模型进行验证分析。主要研究内容包括:1.基于蒙特卡洛方法的一种GaN HEMT小信号等效电路统计模型研究针对由制备过程中的工艺误差所导致的等效电路模型的不准确性,通过测试来自10个批次的56只国产GaN HEMT器件的多偏置电压下的S参数,采用GaN HEMT小信号等效电路经验方程来进行参数提取,提出了GaN HEMT多偏置小信号等效电路参数统计模型。建模方法包括主成分分析法、因子分析法、多元回归模型以及蒙特卡洛方法。应用于GaN HEMT器件的等效电路参数,通过对比原始及所建立模型的均值、标准差、相关矩阵和S参数,并通过分析多偏置电压下的S参数,验证了所建立的小信号等效电路统计模型的准确性。2.基于蒙特卡洛方法的一种GaN HEMT大信号等效电路统计模型研究针对GaN HEMT大信号指标(输出功率Pout、功率附加效率PAE等)的波动性,测试得到10个批次的34只GaN HEMT器件的直流I-V以及大信号测试结果,对非线性漏源电流Ids以及非线性栅电容Cgs、Cgd模型中的参数进行提取。基于小信号等效电路统计模型的建模方法,提出了全参数大信号等效电路统计模型,克服了单一模型不能反映实际电路中的波动范围这一缺点。将得到的大信号统计模型嵌入ADS中进行仿真,通过对比仿真与实测的漏源电流、输出功率和功率附加效率等结果,并通过分析仿真与实测的输出功率和功率附加效率的统计特性,验证了此大信号统计模型的准确性。3.基于响应曲面法的一种GaN HEMT大信号等效电路统计模型研究针对统计模型中由于数据量大而导致电路仿真中的不收敛等问题,基于响应曲面法,并对响应曲面法进行优化改进,提出了一种简便的大信号统计模型。该方法所建立的统计模型数据量较小,具有准确、快速及简便等优点,并且该统计模型在电路仿真中具有很好的收敛性。通过对比仿真与实测的输出功率和功率附加效率等结果,并通过分析仿真与实测的输出功率和功率附加效率的统计特性,验证了通过响应曲面法所建立的大信号统计模型的准确性。4.以统计模型为基础的GaN HEMT高效率功率放大器设计与分析为了验证本文所建立的基于蒙特卡洛方法和响应曲面法的统计模型的准确性,分别以它们为基础,设计了一组Ku波段工作频带为13.7GHz~14.2GHz的E类功率放大器和一组S波段工作频带为2.7GHz~3.5GHz的逆F类功率放大器模块,并对每组功放加工装配三个。通过仿真与实测结果对比显示,实测结果都落在仿真结果的范围内,并与仿真结果的均值非常接近,验证了本文所建立的统计模型的准确性,适合用于电路设计和成品率分析。最后将统计模型应用于一款C波段工作频带为5GHz~6GHz的GaN HEMT功放单片电路(MMIC)中进行仿真,通过与实测结果进行对比分析,进一步验证了此统计模型的准确性,同时也说明此统计模型适用于MMIC功放电路设计以及成品率分析。
[Abstract]:With the rapid development of wireless communication, radar and electronic warfare, the performance of microwave transceivers is higher. As the key component of the transmitter, power amplifier is the most important, the most critical is the power transistor. Its bandwidth, output power, efficiency and working temperature will seriously affect the integrity of the system. Compared with the first generation semiconductor materials represented by silicon (Si) and the second generation of semiconductor materials represented by gallium arsenide (GaAs), the third generation semiconductor material (GaN) has the advantages of large band gap, high saturation electron velocity, high critical breakdown electric field and high thermal conductivity, so it has become an important research in semiconductor devices. The accuracy of the model of.GaN high electron mobility transistor (HEMT) has a great impact on the design of the circuit. At present, some progress has been made in the physical model of GaN HEMT and the large, small signal equivalent circuit model. However, the process error, the model uncertainty, the change of the process parameters and the environmental causes in the production process. The consistency of devices and circuits and the process analysis and optimization design are particularly important. Therefore, it is of great significance to carry out the research on the statistical model of GaN microwave devices. In this paper, a large, small signal equivalent circuit statistical model is established for the GaN HEMT equivalent circuit model, and based on the equivalent circuit statistical model, the design of the equivalent circuit model is designed. The GaN HEMT high efficiency power amplifier is used to verify the statistical model established. The main contents are as follows: 1. a statistical model of GaN HEMT small signal equivalent circuit based on Monte Carlo method is used to study the inaccuracy of the equivalent circuit model caused by the process error in the preparation process, and from the test from 10 batches The S parameters of 56 domestic GaN HEMT devices under multiple bias voltages are obtained by using the empirical equation of the GaN HEMT small signal equivalent circuit to extract the parameters. The statistical model of the equivalent circuit parameters of the GaN HEMT multi biased small signal is proposed. The modeling methods include the principal component analysis, factor analysis, multiple regression model and Monte Carlo method. For the equivalent circuit parameters of GaN HEMT devices, by comparing the mean, standard deviation, correlation matrix and S parameters of the original and established models, and by analyzing the S parameters under the multi bias voltage, the accuracy of the established statistical model of the small signal equivalent circuit is verified by.2., a GaN HEMT large signal equivalent circuit based on the Mont Carlo method. Considering the volatility of GaN HEMT large signal index (output power Pout, power added efficiency PAE and so on), the DC I-V and large signal test results of 34 GaN HEMT devices of 10 batches are tested. The parameters in the nonlinear drain current Ids and the nonlinear gate capacitor Cgs, Cgd model are extracted. The statistical model of the full parameter large signal equivalent circuit is put forward, which overcomes the shortcoming that the single model can not reflect the fluctuation range in the actual circuit. The large signal statistical model is embed in the ADS to simulate the leakage current, the output power and the power added efficiency by comparing the simulated and measured leakage source current. According to the analysis of the statistical characteristics of the output power and power added efficiency of the simulated and measured data, the accuracy of the large signal statistical model is verified by the.3. based GaN HEMT large signal equivalent circuit statistical model based on the response surface method, which is based on the problem of the non convergence in the circuit simulation due to the large amount of data in the statistical model. On the basis of response surface method and the optimization of response surface method, a simple large signal statistical model is proposed. The statistical model of this method has the advantages of small amount of data, accurate, fast and simple, and the statistical model has good convergence in the circuit simulation. The output power of the simulation and the measured is compared. And power added efficiency and so on, and by analyzing the statistical characteristics of the output power and power added efficiency of the simulation and measurement, the design and analysis of the GaN HEMT high efficiency power amplifier based on the statistical model based on the statistical model is verified by the accuracy of the large signal statistical model established by the response surface method to verify the base set up in this paper. Based on the accuracy of the Monte Carlo method and the response surface method, a group of E class power amplifiers with Ku band working band of 13.7GHz~14.2GHz and a group of inverse F power amplifier modules with a group of S band working bands 2.7GHz~3.5GHz are designed, and three units of each group are fabricated. The test results show that the measured results all fall within the range of the simulation results and are very close to the mean of the simulation results. The accuracy of the statistical model established in this paper is verified, and it is suitable for the circuit design and the yield analysis. Finally, the statistical model is applied to a GaN HEMT amplifier monolithic circuit of the C band working band of 5GHz~6GHz. (MMIC) simulation is carried out, and the accuracy of the statistical model is verified by comparison and analysis with the measured results. At the same time, it also shows that the statistical model is suitable for the design of MMIC power amplifier circuit and the analysis of yield.
【学位授予单位】:电子科技大学
【学位级别】:博士
【学位授予年份】:2016
【分类号】:TN386
【相似文献】
相关期刊论文 前10条
1 夏新华,潘志庚;基于统计模型的海水运动仿真[J];计算机仿真;2005年01期
2 赵培东;谢剑英;;图像小波域局部统计模型的拟合优度检验[J];应用概率统计;2008年06期
3 朱青;;Mathematica软件在工业统计模型中的应用研究[J];电子技术;2012年06期
4 曾建军;统计模型的计算机模拟教学系统[J];安徽大学学报(自然科学版);1997年02期
5 董军;徐淼;潘云鹤;;基于统计模型的书法创作模拟[J];计算机学报;2008年07期
6 骆桢轩;;组合统计模型在程序错误定位中的应用[J];计算机工程与设计;2010年19期
7 朱忠林;丁辉;王广志;;利用可变形统计模型进行膝关节建模与运动测量[J];清华大学学报(自然科学版);2013年01期
8 彭嘉雄;张天序;;随机离散图象的统计模型研究[J];自动化学报;1984年01期
9 许小剑,黄培康;一种新的雷达目标散射截面统计模型[J];系统工程与电子技术;1990年02期
10 季玲玲;陈浩;张晔;;基于统计模型预测截断的码率控制算法[J];计算机工程;2008年18期
相关会议论文 前10条
1 阳德青;肖仰华;汪卫;;基于统计模型的社会网络群体关注度的分析与预测[A];NDBC2010第27届中国数据库学术会议论文集(B辑)[C];2010年
2 蒋丹宁;秦勇;刘文;刘义;;基于统计模型的哼唱旋律识别算法[A];第九届全国人机语音通讯学术会议论文集[C];2007年
3 陈永胜;刘小明;任福田;张黔芳;;道路安全设计支持平台之统计模型的研究[A];第七次城市道路与交通工程学术会议论文集[C];2002年
4 严力蛟;全为民;赵晓慧;;千岛湖浮游植物统计模型及其预测[A];浙江省第二届生物多样性保护与可持续发展研讨会论文摘要集[C];2004年
5 刘晶晶;卫水爱;孙俊君;胡海云;邢修三;;生物材料的非平衡统计模型[A];中国力学学会学术大会'2005论文摘要集(上)[C];2005年
6 高山;张艳;徐波;宗成庆;韩兆兵;;基于三元统计模型的汉语分词及标注一体化研究[A];自然语言理解与机器翻译——全国第六届计算语言学联合学术会议论文集[C];2001年
7 彭T3忠谋;;训练用汉语语音资料之音素标示[A];第四届全国人机语音通讯学术会议论文集[C];1996年
8 信飞;孙国武;陈伯民;;自回归统计模型在延伸期预报中的应用[A];第六届长三角气象科技论坛论文集[C];2009年
9 胡振涛;刘先省;;机动目标跟踪的一种改进多传感器数据融合算法[A];第二十四届中国控制会议论文集(下册)[C];2005年
10 张本爱;莫俊永;张玉兰;刘海燕;;任意自旋粒子的一种改进光学统计模型[A];第八届全国核物理会议文摘集(上册)[C];1991年
相关重要报纸文章 前5条
1 葛新权 北京信息科技大学经济管理学院;经济统计模型的建模原则与创新[N];中国社会科学报;2010年
2 李传金;设计新的统计模型[N];经济视点报;2006年
3 沈,
本文编号:2039636
本文链接:https://www.wllwen.com/shoufeilunwen/xxkjbs/2039636.html