請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/85193
標題: | E波段低雜訊放大器與X波段功率放大器之設計 Design of An E-band Low Noise Amplifier Using 28-nm CMOS process and A X-band Power Amplifier Using 0.25-um GaN HEMT process |
作者: | Chuan-Li Chung 鍾傳立 |
指導教授: | 黃天偉 (Tian-Wei Huang) 黃天偉 (Tian-Wei Huang | tihuang@ntu.edu.tw | ), |
關鍵字: | 互補式金屬氧化物半導體,低雜訊放大器,氮化鎵,功率放大器,寬頻,E波段,X波段,增益平坦度,相位陣列, CMOS,Low noise amplifier,GaN,Power amplifier,Wideband,E-band,X-band,Gain flatness,Phased array, |
出版年 : | 2022 |
學位: | 碩士 |
摘要: | 本論文包含三個部份。第一部分是一個針對天文應用之次世代超大型陣列(ngVLA)需求所設計的E波段(E-band)的低雜訊放大器,使用28奈米互補式金氧半場效電晶體(CMOS)製程。第二部分是一個針對相位陣列系統之需求所設計的X波段(X-band)功率放大器,使用0.25微米氮化鎵高電子移動率電晶體(HEMT)製程製作。最後討論利用高通濾波器來增加整體電路穩定性,以及在偏壓電路(bias circuit) 模擬上討論多種方法與組合來增強電路低頻穩定度。 第一部分提出使用28奈米HPC-plus CMOS製作的E頻帶擁有 3.8 dB 雜訊指數的低雜訊放大器,其電路由一級共源極(common-source)架構串聯兩級疊接放大器(cascode amplifier),當中在疊接放大器(cascode amplifer)的共閘極(common-gate)使用增益提升(gm-boosting)技術,能使整體增益明顯上升,並在疊接放大器之共源、共閘極間串接傳輸線將輸出阻抗移至匹配損耗低之區域。整體電路多大數使用傳輸線作為匹配網路來達到高平坦增益寬頻的低雜訊放大器,分段式的匹配使各級偏壓調整,可調整整體增益;此電路在 93.4 GHz 達到小訊號增益最大值 22.3 dB。透過偏壓調整,3-dB 頻段內的小訊號增益都能達到 16 dB 以上,且整體增益平坦度達± 0.5 dB 內,3-dB 小訊號頻寬為 63.8 到 93.1 GHz,並在 68 GHz 達到最小 3.8 dB 的雜訊指數,頻帶內雜訊指數平均為 5.24 dB;直流功耗約為 42 mW,而晶片的總面積約為 0.5 平方毫米。 第二部分呈現一顆使用 0.25 微米氮化鎵高速場效電晶體(HEMT)製程製作的31.3 dB 輸出功率放大器,這顆功率放大器的操作頻率鎖定在 9-11 GHz 來符合X波段的應用。我們採用了兩極共源級放大器來維持良好的增益,選擇深Class-AB類之偏壓情況來獲取最高的效率。再者,我們也使用了電阻並聯電容、共模穩定電阻來增強電路穩定度。此外,考慮製程變異,將小電容分別改為串聯大電容實現減少變異影響。量測結果顯示此放大器在 9.2 GHz 到 11.5 GHz 頻寬中,提供約 30 dBm 的飽和輸出功率(Psat)和最高30%的功率附加效率(PAE)。 最後一部分利用偏壓電路的模擬與高通濾波器的設計增強電路穩定度,以第一部分與第二部分之量測結果進行研究,詳細的說明如何解決低頻震盪、偏壓電路與震盪關係的現象討論。在低頻率(0.1 GHz)以下,電路內之偏壓電路難以設計並顧全低頻,時常在量測時發生震盪訊號。最後一部分透過 Advanced Design System(ADS)軟體模擬出潛在的震盪風險,利用 ADS 模擬比對前兩部分電路之量測結果,歸納出合適的檢驗及解震方法。 This thesis consists of three parts. The first part is an E-band low-noise amplifier with a minimum noise figure of 3.8 dB designed for astronomical applications, especially in next generation very large array (ngVLA), using 28-nm CMOS process. The second part is a X-band power amplifier designed for phased array system applications using 0.25 um GaN HEMT process. Finally, we discuss the use of high-pass filters to improve circuit low-frequency stability as well as a number of techniques and combinations to improve circuit overall stability. The first part proposes an E-band low-noise amplifier using 28-nm CMOS HPC-plus with a minimum noise figure of 3.8 dB, with a one-stage common-source structure followed by a two-stage cascode amplifier. The gain-boosting technique is used at the common-gate of the cascode amplifier to increase the overall gain significantly, and the output impedance is matched to a low-loss region by connecting transmission lines in series between the common-source and common-gate of the cascode amplifier. The overall circuit mostly uses transmission lines as a matching network to achieve a high gain flatness wideband low-noise amplifier, and the segmented matching enables bias adjustment at each stage to adjust the overall gain. The circuit achieves a maximum small-signal gain of 22.3 dB at 93.4 GHz. With bias adjustment, the small-signal gain in the 3-dB bandwidth can reach more than 16 dB, and the overall gain flatness is within ±0.5 dB. The 3-dB small-signal bandwidth is 63.8 to 93.1 GHz, and minimum noise figure of 3.8 dB is achieved at 68 GHz, with an average in-band noise figure of 5.24 dB. The total area of the chip is about 0.5 mm2. In the second part, we present a high output power amplifier using 0.25-um GaN HEMT process, the operating frequencies of this work is targeted at 9-11 GHz for X-band applications. In order to obtain good gain performance, a two-stage common source amplifier was chosen for the design and deep Class-AB bias was selected to obtain the highest efficiency. In addition, we use a resistor in parallel with the capacitor and common mode resistors to improve the stability of the circuit. Moreover, the effect of variation can be reduced by considering process variation using large capacitors connected in series instead of small capacitors. Measurements show that the amplifier provides a saturation output power (Psat) of about 30 dBm and a peak power added efficiency (PAE) of up to 30% over a bandwidth of 9.2 GHz to 11.5 GHz. The following sections discuss the use of high-pass filters to improve overall circuit stability, as well as a variety of methods and combinations for bias circuit simulation to enhance circuit low-frequency stability. The phenomenon of bias circuit in relation to oscillation is also discussed. At low frequencies (0.1 GHz) and below, which are difficult to simulate with EM simulation software, oscillation signals are often present during measurements. The last part is simulated by the Advanced Design System (ADS) software to identify potential oscillation risks, and the ADS simulation is used to compare the measurement results of the first two parts of the circuit and to summarize the appropriate inspection and solving oscillation methods. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/85193 |
DOI: | 10.6342/NTU202200070 |
全文授權: | 同意授權(限校園內公開) |
電子全文公開日期: | 2027-08-08 |
顯示於系所單位: | 電信工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-1601202214134000.pdf 目前未授權公開取用 | 7.63 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。