Class-D放大器(D類放大器)衍生的目的在於改善及提升AB類放大器(Class-AB)的特性，簡述如下：
1.	效能提升，可達>85%(傳統的AB類放大器效能僅為50%)。
2.	降低散熱片的倚賴性，甚至在低功率(20W以下)時可完全將散熱片捨棄，因此商業用途的產品體積即可縮小，如：手機、PMP等。
3.	但同時也造成另一個問題的出現：EMI的干擾。
簡單說，Class-D放大器內部電路的設計方式是將原始的類比信號波形，與比它更高頻率的三角波（或鋸齒波）進行電壓比較（透過電壓比較器），如此便可將以振幅高低性表示的信號調變成以脈波寬窄性表示的信號，此即是脈寬調變（Pulse Width Modulation；PWM），之後將PWM信號輸出到MOSFET場效電晶體上的閘極，以控制電晶體的導通、關閉(簡稱：開關頻率)，同時也在這個階段進行信號功率放大，最後MOSFET的輸出端連接LC（電感、電容）低通濾波電路，將PWM的載波濾除，使原始信號波形重新呈現。
因此，此份研究報告的內容將針對Class-D放大器對AM/FM的干擾與排除進行研究與分析。究其干擾原因，CLASS-D放大器的開關訊號會經由低通濾波器的電感輻射出去，對調幅(AM/FM)廣播產生干擾，被干擾的電台主要是位在CLASS-D放大器開關頻率的諧波上。如原本的開關頻率是300KHz， 則在其倍數頻率上的電台(600KHz、900KHz、1200KHz、1500KHz)通通會被影響，主要症狀是倍頻處的電台接收靈敏度突然變差或遭受到蓋台。因此，研究的重點在於干擾的產生及如何排除的方式，其概念如下簡述：原理是產生一個外部的基準時脈給Class-D放大器，利用IC內部PWM開關頻率的改變抑止倍頻干擾而產生的問題。如：干擾AM電台問題，將設有兩個震盪頻率可供選擇(430KHz與480KHz)，在接收不同電台時可以切換系統震盪頻率來避掉干擾. 一般當接收電台的頻率位在CLASS-D放大器開關頻率倍頻的±30KHz以內時就可以更換系統主震盪頻率，每次只有一個震盪器在工作，頻率的選擇由外部電路來決定，以解決倍頻的干擾。FM的干擾解決方式採用金屬材質的外殼屏蔽，以便衰減諧波輻射的干擾能量，進而使得FM接收正常。

Only the power amplification of A, B, mostly used in the amplifier of audio systems had learned in school. And most consumer audio and public announcement systems use Class-AB topology today. However how is Class D developed?
   In fact, D amplifier (Class-D) is to improve and promote the performance of AB amplifier (Class-AB), and short depictions as followed:
1.	Increasing efficiency, can reach > 85% (the traditional Class-AB amplifier is only 50% efficient.)
2.	Decreasing the dependence of the heat sink, even in the low power (under 20W) the heat sink is not required, so the size of products for the commercial uses can be diminished, for instance: Mobile phone, PMP, etc.
3.	But it could cause another problem at the same time: interference with EMI.
   Briefly, the design of Class-D’s circuits is to compare the original analog signal waveform with triangular waveform or sawtooth waveform operated at the high frequency. So that the voltage level of signal transmits to the pulse width, this is so called Pulse Width Modulation; PWM. Then PWM transmits to the transistor of MOSFET to control the switch of transistor (short for: Switching Frequency), at the same time power is amplified. Finally the output connecting MOSFET of LC (inductance, capacitor) drives through low pass circuits and filters through the carrier wave to reveal the original wave of single again. 
   Therefore this is the research of interfere with AM/FM and elimination by Class D.
The cause of interference is that the switch signal of Class-D will go out via the low pass inductance radiation and then Class-D is interfered with the amplitude modulation (AM/FM). The radio station interfered is on the distortion waves of the switch frequency. If the original signal having a switch frequency of 300 KHz, for example, the radio stations broadcasting on the multiple frequency (600 KHz,900 KHz,1200 KHz,1500 KHz) are affected. The main symptoms are that the delicacy of receiving the signal by the radio station becomes weak suddenly or one radio station is covered with the other one.
   Thus, the key point of this research is finding out the methods to avoid from interference and estimations. The concept is stated as below. To install an outside basic clock in Class-D and use the change of the switch frequency of PWM to avoid the interference with multiple frequency. Take the interference with AM for example; there are two choices of basic clocks (430 KHz and 480 KHz). The way to avoid interference is by exchanging frequency automatically. Generally the system will change the basic clock to a tolerance of +/- 30 KHz. There is always one basic clock operated. The choice of frequency will depend on the outside circuits in order to solve the problem of multiple frequencies.

第一章  緒論 1
1.1 研究動機 1
1.2 Class-D放大器的演化過程 3
1.2.1  Class-AB放大器與Class-D放大器之兩者主要差異 3
1.2.2  A類功率放大器之介紹  4
1.2.3  B類功率放大器之介紹	6
1.2.4  AB類功率放大器之介紹	9
第二章  Class-D放大器的內部電路設計架構及原理探討 10      
2.1 Class-D放大器之介紹 	10
2.2 Class-D放大器之功率效應	12
2.3 Class-D放大器的動作原理	15
2.3.1 Class-D 放大器IC概念	17
2.3.2 Class-D 放大器IC應用電路 19
2.3.3 PWM動作原理  22
2.3.4低通濾波器原理及電路	24
第三章  Class-D放大器對AM/FM的干擾源的產生與分析  29
3.1  Class-D放大器對AM/FM干擾產生的現象29                   
3.1.1測試架構圖 30
3.1.2 PWM對於AM/FM之干擾概述 31
3.1.3針對干擾源現象之測試模式 31
3.1.4測試後之波形 31
3.2  干擾源產生的原理 36
3.3  對於AM/FM的干擾現象解析 37
第四章  對於倍頻及EMI干擾的排除與模擬實驗結果之探討  40
4.1  排除倍頻/EMI干擾之電路架構及原理解說 40
4.1.1 排除倍頻干擾之應用電路圖(2顆Class-D IC)	40
4.1.2 CLOCK SOURCE SEL實體PCB板 41
4.1.3 排除倍頻干擾之原理解說(AM之干擾解決方法) 41
4.1.4 排除EMI干擾之原理解說(FM之干擾解決方法)	42
4.2  電路模擬實驗結果  43
4.2.1 AM干擾解決方式之實驗結果 43
4.2.2 FM干擾解決方式之實驗結果 50
第五章  結論與未來展望 51
Reference	 52

 
圖目
圖 1.1: A類放大器輸出架構圖	4
圖 1.2: A類放大器的特性曲線	5
圖 1.3: B類放大器輸出架構圖	6
圖 1.4: B類放大器的特性曲線	7
圖 1.5:交越失真造成說明曲線	8
圖1.6：AB類放大器輸出架構圖…………………………………………………………9
圖 2.1:傳統(A、B、AB)類放大器最後一級訊號之關係	11
圖 2.2:數位化-D類放大器最後一級訊號之關係	11
圖 2.3: Class-D放大器之輸出功率示意圖	13
圖 2.4: Class-D放大器內部邏輯電路圖及外部LC之示意圖	16
圖 2.5: Class-D放大器 IC 方塊圖(僅為單聲道示意圖)	17
圖 2.6: Class-D 放大器IC 應用電路圖	19
圖 2.7:無輸入訊號，IC輸出端(低通濾波電路(LC電路)前)	20
圖 2.8:輸入1KHz訊號，輸出端(低通濾波電路(LC電路)前)	20
圖 2.9:輸入1KHz訊號，輸出至喇叭端的音源波形(低通濾波電路(LC電路)後)….	21
圖 2.10: PWM的輸出波形示意圖….	23
圖 2.11:低通濾波器的頻率響應….	25
圖 2.12:特沃斯 (Butterworth)、卻比雪夫(Chebyshev)、貝塞爾(Bessel) 濾波器的特性 曲線….	26
圖 2.13:二階LC型的巴特沃斯(Butterworth)濾波器….	26
圖 2.14:實際的BTL電路….	27
圖 3.1: Class-D放大器 IC應用於汽車音響上的PCB板………………………..…….29
圖 3.2:測試架構圖	30
圖 3.3: AM：Class-D放大器“不動作”時之環境測試	32
圖 3.4: AM：Class-D放大器“動作”時之環境測試，基頻270KHz	32
圖 3.5: AM：Class-D放大器“動作”時之環境測試，二次諧波600KHz	33
圖 3.6: AM：Class-D放大器“動作”時之環境測試，三次諧波900KHz	33
圖 3.7: AM：Class-D放大器“動作”時之環境測試，四次諧波1200KHz	34
圖 3.8: FM：Class-D放大器“不動作”時之環境測試………………………………..	34
圖 3.9: FM：Class-D放大器“動作”時之環境測試	35
圖 3.10: 關關頻率之波形	36
圖 4.1: AM干擾解決方式之電路圖	40
圖 4.2: CLOCK SOURCE SEL實體PCB板	41
圖 4.3: FM之解決方式	42
圖 4.4: 頻率430KHz時的模擬實驗波形	45
圖 4.5: 頻率480KHz時的模擬實驗波形…………………………………………..…...47圖 4.6: 頻率430KHz與480KHz時的模擬實驗之關係波形…………………………..49
圖 4.7: FM模擬實驗結果的波形                                           50





表目
表 2.1：對應特定fc和R的電感（L）值和電容（C）值	28
表 3.1: 測試儀器資料                                                    30
表 3.2: AM的頻率表...…………….……....………………………..................................37
表 4.1: AM頻道對主震盪頻率模擬實驗結果之數據關係表….....……….....................43

[1] Nonaka, S., K. Kesamaru, K. Yamasaki and M. Nishi,” Interconnection system with single phase IGBT PWM CSI between photovoltaic arrays and the utility line”,IEEE Industry Applications society, vol.2(10), 1990,pp.1302-1307.
[2] NONAKA, S. and Y. Neba ,”Single phase PWM current source converter with double-frequency parallel resonance circuit for DC smoothing”, IEEE Industry Applications society, vol.2(10),1993, pp.1144-1151.
[3] Nielsen, K, “A review and comparison of pulse width modulation (PWM) methods for analog and digital input switching power amplifiers”, Presented at the 102nd Audio Engineering Society Convention, Munich, Germany, Preprint 4446,1997.
[4] Mosely, I. D., “Effect of dead time on harmonic distortion in class-D audio amplifiers”, IEEE Electronics Letters, vol.35(12),May 1999,pp. 950-952.
[5] Mellor, P.H.; Leigh, S.P. and Cheetham, B.M.G, “Improved sampling process for a digital, pulse-width modulated, class D power amplifier”, IEEE Colloquium on Digital Audio Signal Processing, pp. 3/1-3/5, 1991.
[6] Sokal, N O & Zulinski, R E, ”Class D high-efficiency switching mode tuned power amplifier with oly one inductor and one capacitor in load network approximate analysis”, IEEE Trans, Solid-State Circuits, vol. SC-16, August 1981, pp. 380-387.
[7] M. K. Kazimierczuk, “ Class D voltage-switching MOSFET power amplifier ” IEEE PROCEEDINGS-B, vol.138, No.6, Nov. 1991.
[8] Attwood, B.E., “Design parameters important for the optimization of very-high-fidelity PWM (class-D) audio amplifiers”, Journal of the Audio Engineering Society, vol 31, No 11, November 1983, pp. 842-853.
[9] Adel S.Sedra and Kenneth C.Smith., Microelectronic Circuits - 3rd ed. , 1991.
[10] S.Franco, Design with Operational Amaplifers and Analog Integrated Circuits, 1988. 
[11] K.Lacanette, ed., Switched Capacitor Filter Handbook, National Semiconductor, April 1985. 
[12] G. Temes & S. Mitra, Modern filter Theory and Design, Wiley, 1973. 
[13] M. Steffes, Simplified Component Value Pre-Distortion for High Speed Active   Filters OA-21, Comlinear, March 1993. 
[14] A.B. Williams, F.J. Taylo, Electronic Filter Design handbook, 1988. 
[15] Dijkmans, E., C. Dijkmans and J. A. T. M. van den Homberg, ”PWM amplifier with feedback loop integrator”, U.S. Patent 6,300,825,2001.
[16] Garverick, et al., “High frequency pulse width modulation driver, particularly useful for electrostatically actuated MEMS array,” US Patent 6,543,286, April 8, 2003.
[17] www.princeton.com.tw
[18] www.maxim-ic.com
[19] www.ti.com
[20] www.national.com