砷化鎵與氮化鎵之Doherty功率放大器研究

Abstract

在無線通訊系統中，微波訊號在空氣中傳送時，因受雜訊干擾與傳遞時訊號逐漸減弱影響，常導致接收機收到訊號失真而解碼錯誤，因此若能加強傳送訊號使其具有較好的訊號與雜訊能量比，則能增加解碼正確的機率，故在通訊系統中發射機的功率放大器如何具有更大的輸出功率便是重要的課題。除了高功率輸出之外，更要求高效率的操作。由於功率放大器在發射機中消耗了主要部的的能量，因此將效率最佳化是另一個重要的關鍵。 在本篇論文中提出了三個功率放大器，分別為38 GHz砷化鎵功率放大器、38 GHz類多爾蒂砷化鎵功率放大器與10 GHz 氮化鎵多爾蒂功率放大器。 在本論文第二章中，我們提出了一個在38 GHz使用0.15μm GaAs的功率放大器來實現高功率和高增益。功率放大器由兩級共源級結構組成。然而功率級和驅動級的電晶體尺寸選擇都比傳統設計大。但是，模擬結果依然實現了高效率和增益。在測量中，由於缺乏熱考慮，S參數和功率性能低於模擬結果。 在第三章中，提出了一個10 GHz使用0.25μm氮化鎵的 Doherty功率放大器來提高回推效率。主路徑由共源級結構組成，偏壓在AB類，以在小信號時提供足夠的增益。輔助路徑也由共源級結構組成，並且偏壓在C類，為了在大信號時提供高增益，並在小信號操作時節省直流功耗。此外本章還對功率放大器模組進行了模擬與量測。然而輸出功率是瓦特等級，故設計應該注重熱效應和測量環境。所以我們盡可能地把環境建立的越低溫。 在第四章中，提出了一個38 GHz使用0.15μm砷化鎵的類Doherty功率放大器來提高回推效率。對稱DPA架構的主要缺點是在毫米波頻段（MMW）下的C類峰值放大器的低電流。因此主路徑和輔助路徑都由共源級組成，且分別偏壓在A類和AB類。與第二章設計方法不同的是各個放大器的負載阻抗和偏壓狀況選擇不同，以及主路徑的補償線。此外輸入反射係數與模擬結果不一致，附上了偵錯結果。

In wireless communication system, when the microwave signal transmits in the air, the receiver could encounter the signal distortion which causes the decoding error due to the noise interference. Therefore, if we can strengthen the transmitting signal and make better signal-to-noise ratio (SNR) of the signal, we can improve the propability of correct decoding for the reason that how to make the power amplifier deliver more output power in wireless transmitter. Besides the high output power, the demand for high efficiency of power amplifier in wireless communication system has increased because power amplifier consumes the most dc power in the transmitter. Therefore, the optimization of the efficiency becomes a key issue. In this thesis, a 38 GHz power amplifier, a 38 GHz quasi-Doherty power amplifier in 0.15-μm depletion mode GaAs pHEMT MMIC process and a 10 GHz 0.25-μm GaN HEMT Doherty power amplifier are proposed. In chapter 2, a 38 GHz power amplifier in 0.15-μm GaAs pHEMT process is proposed to achieve high power and gain. The proposed PA consists of two stage common-source architecture. However, the device size of the power stage and the driver stage are both selected larger than the traditional design. However, the simulated results also achieve high efficiency and gain. In the measurement, due to the lack of thermal consideration, the S-parameter and power performance is lower than the simulated results. In chapter 3, a 10 GHz 0.25-μm GaN HEMT Doherty power amplifier is proposed to improve the back-off efficiency. The main path consists of common source topology and is biased at class-AB to provide enough gain at small-signal. The auxiliary path also consists of common source topology and is biased at class-C to provide high gain at large-signal and to save dc power consumption for small-signal operation. In addition, the PA module is also simulated and measured in this chapter. However, the output power is the watt-level, so the design should be cared about the thermal effect and measured surrounding. Therefore, we set up the environment as cool as possible. In chapter 4, a 38 GHz quasi-Doherty power amplifier in 0.15-μm is used to improve the back-off efficiency. The major drawbacks of the symmetrical DPA architecture is the low current of the class-C peak amplifier at millimeter wave frequency (MMW). Therefore, the main path and auxiliary path both consist of common source topology, which are biased at class-A and class-AB, respectively. The different to chapter 2 are the load impedance and the bias condition choice of each amplifier and the compensation line of the main path. Besides, the input return loss is different from simulated result. The debug result is provided.