502494 A7 B7 五、發明説明(1 ) 發明背景 本發明係關於不同態樣的射頻(R F )通訊系統。 (請先閱讀背面之注意事項再填寫本頁) 諸如用於語音及資料之蜂巢式電話等無線通訊已相當 流行。目前,使用數種無線設計,包含G S Μ、T D Μ A 、及CDMA。在這些設計中,CDMA顯然逐漸變成美 國、歐洲及亞洲市場的標準。因此,C D Μ A蜂巢式電話 或基地台中用於產生R F訊號之功率線性放大器的效率及 耗電相當重要。由於幾個理由,使用低效率線性放大器是 有害的。 這些放大器會耗掉可觀數量的能量,特別是在以電池 操作的蜂巢式電話中,這是有問題的。在基地台中,耗電 量也是問題。由基地台中很多低效率的放大器所產生的熱 量會促使元件故障,因而降低可靠度。功率放大器的線性 也是重要的。在同時傳送多個訊號之基地台中,具有不佳 的線性之放大器會造成這些訊號無意的混合。 經濟部智慧財產局員工消費合作社印製 在R F通訊系統中可使用一些放大器等級型式,包含 等級A、等級A B、等級C、等級E、等級F、及等級D (有時稱爲數位放大器)。但是,當在R F範圍中操作時 ,每一這些型式的放大器具有顯著的問題。舉例而言,等 級A及等級A B放大器具有非常差的效率。雖然相較於等 級A及等級A B型放大器,等級C、E、F及D放大器具 有改進的功率效率,但是,它們不適用於線性應用。此外 ,等級E放大器在其輸出會遭遇嚴重的超調問題,因而限 制其用途。等級F放大器在重覆的輸入訊號下,呈現相當 本紙張尺度適用中國國家標準(CNS ) A4規格(21〇X297公釐) -4- 502494 - A7 ______B7 五、發明説明(2 ) 良好的輸出切換特性。但是,在諸如蜂巢式電話或基地台 中經常遇到的非重覆輸入訊號下,由諧波造成的問題會變 嚴重。結果,等級F放大器難以用於R fT線性功率放大器 應用中。 傳統的等級D放大器在較低頻率時具有線性操作特性 且通常有高效率,但是在較高頻率時,以往均會有數個缺 點。最嚴重的是,在諸如RF等較高頻率時,它們在其輸 出電晶體處出現切換問題。當這些電晶體快速地開及關時 ,在輸出處發生包含高位準的電流及電壓之切換過渡,造 成超調及下越。 圖1係顯示根據本發明的特定實施例之R F帶通雜訊 頻譜成型放大器1 0 0。重要的是要注意,在背景說明中 包含此新穎的放大器配置僅爲了提供與此處所述的此基本 配置之進一步改良有關的背景資訊。不應將此配置假定爲 習知技藝。事實上,放大器1 0 0代表不同於舊有的帶通 結構之新穎結構。 RF放大器100包含選頻網路1〇其使用連續 時間回饋,以將經過調變的R F輸入施以雜訊頻譜成型。 網路1 0 2包括至少一共振器級,此共振器級具有設計成 通過以諸如9 0 ΟΜΗ z爲中心的頻帶之轉換函數。 A / D轉換器1 0 4會使用取樣頻率f s (舉例而言, 在本實施例中爲3 · 6 G Η z ),將經過雜訊頻譜成型的 RF訊號轉換成數位資料。A/D轉換器1 〇 4包括單一 比較器。或者,A/D轉換器104包括用以實施三位準 本紙張尺度適用中國國家標準(CNS ) Α4規格(2丨〇><297公釐) (請先閱讀背面之注意事項再填寫本頁) 、1Τ -Φ-. 經濟部智慧財產局員工消費合作社印製 -5- 502494 A7 B7 五、發明説明(3 ) 切換之二比較器。 (請先閲讀背面之注意事項再填寫本頁) 閘極驅動電路1 0 6取用來自A/D轉換器1 0 4之 脈沖列並產生閘極驅動以用於放大器1 〇 〇的功率輸出級 之每一 FET 108和1 10。所示的輸出功率級包含 三電感LI、L2及L3、以及電容器C1。當FET 1 0 8及1 1 0中的對應者閉接時,此配置分別在節點A 和B處產生分別的共振。 經由回饋路徑1 1 2,提供回饋至選頻網路1 0 2之 連續時間回饋。功率級的輸出訊號會傳送給匹配網路 1 1 4,匹配網路1 1 4會將輸出的R F訊號傳送給用於 傳輸之天線1 1 6。 頂部及底部F E T 1 0 8和1 1 0會由藉由閘極驅 動電路1 0 6於A / D轉換器1 0 4的輸出產生之閘極訊 號交錯地驅動。假使F E T 1 0 8和1 1 0由諸如方波 等切換訊號驅動時,每一裝置之寄生的閘極電容會造成嚴 重及無謂的損失,此損失與閘極訊號的頻率、閘極電容、 經濟部智慧財產局員工消費合作社印製 及輸入電壓平方成正比,亦即f cv2損失。因此,電感器 1 1 8會與每一 F E T閘極串聯以共振消除f c v 2損失。 如上所述,選頻網路1 0 2配合連續時間回饋路徑1 1 2 ,提供雜訊頻譜成型給放大器。網路1 0 2包括一或更多 共振器級。 雖然這將減少這些無謂的損失,但是,仍有其它不希 望的結果。首先,會在閘極訊號過渡時於每一 F E T的輸 入處產生不需要的且實質超調(或下越)電壓。其次,每 本紙張尺度適用中國國家標準(CNS ) A4規格(210 X 297公釐) -6 - 502494 A7 B7 五、發明説明(4 ) 一閘極過渡會在每一閘極產生實際上要無限地連續之共振 。假使放大器僅用於產生0與1交錯之規律樣式(頻率及 相位調變應用),則由於會造成所需的位元樣式,所以, 此共振不成問題(假使可以接受超調/下越)。對應需要 非週期性的位元樣式之應用而言,此共振干擾所需的位元 樣式之產生。 此外,使用固定的時計以產生用於A / D轉換器 1 0 4之時計(亦即f s )具有缺點。首先,在節點A與 B處的共振之時序通常需要符合此時計的時序,以使效率 最大。但是,事實上,舉例而言,由於來自匹配網路 1 1 4及處理變異之反應,節點A及B處的共振傾向於在 頻率上變動。其次,由於設計複雜度,通常不希望在此設 計中使A / D轉換器需要分別的時計。 如同所知,圖1之放大器配置允許二量化狀態。藉由 二量化狀態,會有較多數目之導致高驅動損失的訊號過渡 〇 根據脈寬調變(p w Μ )及其它調變技術,必須有效 地處理切換及輸出濾波器級所造成的延遲,以減輕此延遲 在電路穩定上造成的不利效果。對於具有相當高功率切換 級之調變器而言,由於延遲相對於迴路之脈沖重覆頻率會 變得很大,所以,此點特別真實。將參考圖2及3,說明 傳統上對此延遲處理的解決之道。 圖2係典型的調變器迴路2 0 0之方塊圖。調變器 2 0 2的邏輯輸出會驅動反相功率級2 0 4,反相功率級 本紙張尺度適用中國國家標準(CNS ) Α4規格(210><297公釐) (請先閲讀背面之注意事項再填寫本頁) 訂 Φ·. 經濟部智慧財產局員工消費合作社印製 502494 A7 B7 五、發明説明(5 ) (請先閱讀背面之注意事項再填寫本頁) 2 0 4的輸出會由輸出濾波器2 0 6濾波。回饋電阻器 2 0 8及衰減電阻器2 1 2係用於從調變器2 0 2的迴路 輸出導入負回饋。圖3係顯示來自圖2的無延遲補償的調 變器迴路之二波形3 0 2和3 0 4。如同所示,由於切換 及濾波器級導入的延遲,波形3 0 4的正擺幅(亦即,輸 出濾波器2 0 6 )與波形3 0 2的正擺幅(亦即,調變器 202的邏輯輸出)幾乎同相。 因此,在迴路的原始設計考慮負回饋之情形中,延遲 會將其轉換成正回饋,因此造成迴路不穩定(未顯示)之 結果。因此,設置與回饋電阻器2 0 8並聯之濾波電容器 2 1 0以用於延遲補償。電容器2 1 0在回饋回路中會產 生零,對於高頻而言,可將電阻器2 0 8和2 1 2所造成 的衰減旁通。 經濟部智慧財產局員工消費合作社印製 不幸地,由於在延遲導入迴路之後執行此型式的延遲 補償,所以難以校正與迴路穩定度有關之所有的延遲負結 果。事實上,此型的補償技術僅有有限的成果,結果,限 制了現今調變器之延遲容許度及整體性能。一解決之道係 使用直接來自調變器的邏輯輸出之回饋訊號(亦即,在延 遲被導入之前)結合功率級及/或輸出濾波器的輸出。不 幸地,雖然調變器(3 0 2 )的輸出與濾波器輸出( 3 〇 4 )具有類似的特徵,但是,由於功率級的非線性本 質,所以,有用的調變器迴路頻率範圍的內容會有顯著的 差異。這將於回饋這些訊號時難以取得高逼真操作。 基於上述,需要克服上述不同問題之帶通放大器。 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -8- 502494 A7 B7 五、發明説明(6 ) 發明槪述 (请先閱讀背面之注意事項存填寫本買} 根據本發明的一態樣,提供在回饋迴路中具有選頻網 路之帶通放大器。類比對數位轉換器耦合至選頻網路。切 換級耦合至類比對數位轉換器以產生連續時間輸出訊號。 切換級包含至少一共振電路,其係以共振頻率共振並藉以 產生至少一部份連續時間輸出訊號。連續時間回饋路徑會 感測及回饋連續時間輸出訊號給選頻網路。 也設置用於帶通放大器之切換級。切換級在半橋配置 中包含第一和第二切換開關,第一及第二切換開關均具有 與其相關之寄生電容。切換級又包括多個電感器。二共振 電路係由寄生電容及電感器形成,二共振電路係以共振頻 率交錯地共振。 經濟部智慧財產局員工消費合作社印製 根據本發明的另一態樣,由於這些電感器需要在F E T閘極減少f c v 2的損失,所以,本發明提供閘極驅動電 路控制閘極訊號以克服超調及共振問題之技術。亦即,閘 極驅動電路所產生的閘極脈沖之領先及拖後緣會暫時被引 至邏輯位準、帶回至原始邏輯位準、然後一旦在F E T閘 極的訊號位準安定在所需位準時會被帶回至所需的邏輯位 準。一般而言,本發明使用控制邏輯及電感器以將閘極電 容器的邏輯狀態充電。 因此,本發明提供用於將切換開關從第一邏輯狀態切 換至第二邏輯狀態之方法及裝置。切換開關具有閘極端, 閘極端具有閘極電容及與其相關之電阻。根據不同實施例 本紙張尺度適用中國國家標準(CNS ) A4規枱'.(210X297公釐) -9 - 502494 A7 __B7_ 五、發明説明(7 ) (請先閱讀背面之注意事項再填寫本頁) ,此電阻可爲切換開關的輸入電阻與外部串聯及並聯的電 阻器之任意組合。閘極端也具有耦合至其之串聯電感。當 切換開關處於第一邏輯狀態中時,對應於第二邏輯之脈沖 會施加至電感。脈沖具有與其相關之第一位準的能量,此 能量高至足以克服電阻造成的衰減(因此允許閘極端達到 對應於第二邏輯狀態的訊號位準,藉以將切換開關切換至 第二邏輯狀態),且低至足以減緩導因於電感及閘極電容 之振盪(以致於在切換開關接著過渡至第一邏輯狀態之前 ,閘極端安定於訊號位準)。 根據本發明的另一態樣,使用放大器設計的輸出切換 級中之一或更多共振以產生用於放大器的A/D轉換器之 時計訊號。根據一實施例,在輸出級包括二切換裝置的情 形中,交錯地使用與二裝置有關之共振節點以產生時計訊502494 A7 B7 V. Description of the invention (1) Background of the invention The present invention relates to radio frequency (R F) communication systems in different aspects. (Please read the notes on the back before filling out this page.) Wireless communications such as cellular phones for voice and data have become quite popular. Currently, several wireless designs are used, including G S M, T D M A, and CDMA. In these designs, CDMA is clearly becoming the standard for the US, European, and Asian markets. Therefore, the efficiency and power consumption of a power linear amplifier used to generate RF signals in a CDM cellular phone or base station are very important. The use of low-efficiency linear amplifiers is harmful for several reasons. These amplifiers consume considerable amounts of energy, which is problematic especially in battery-operated cellular telephones. In base stations, power consumption is also an issue. The heat generated by many inefficient amplifiers in the base station can cause component failures, reducing reliability. The linearity of the power amplifier is also important. In base stations transmitting multiple signals at the same time, amplifiers with poor linearity can cause these signals to inadvertently mix. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs There are several amplifier grades that can be used in RF communication systems, including grade A, grade A B, grade C, grade E, grade F, and grade D (sometimes called digital amplifiers). However, each of these types of amplifiers has significant problems when operating in the R F range. For example, Class A and Class A B amplifiers have very poor efficiency. Although Class C, E, F, and D amplifiers have improved power efficiency compared to Class A and Class A B amplifiers, they are not suitable for linear applications. In addition, Class E amplifiers experience severe overshoot problems at their outputs, limiting their use. Under repeated input signals, the grade F amplifier presents the paper size corresponding to the Chinese National Standard (CNS) A4 specification (21 × 297 mm) -4- 502494-A7 ______B7 V. Description of the invention (2) Good output switching characteristic. However, the problems caused by harmonics can become serious with non-repetitive input signals such as those found in cellular phones or base stations. As a result, Class F amplifiers are difficult to use in R fT linear power amplifier applications. Traditional grade D amplifiers have linear operating characteristics at lower frequencies and usually have high efficiency, but at higher frequencies, there have been several shortcomings in the past. Most seriously, they have switching problems at their output transistors at higher frequencies such as RF. When these transistors are turned on and off quickly, a switching transition involving high levels of current and voltage occurs at the output, causing overshoot and overshoot. FIG. 1 shows an RF band-pass noise spectrum shaping amplifier 100 according to a specific embodiment of the present invention. It is important to note that the inclusion of this novel amplifier configuration in the background note is only to provide background information regarding further improvements to this basic configuration described here. This configuration should not be assumed to be a know-how. In fact, the amplifier 100 represents a novel structure different from the old band-pass structure. The RF amplifier 100 includes a frequency selective network 10 which uses continuous time feedback to apply a modulated RF input to the noise spectrum shaping. The network 102 includes at least one resonator stage having a transfer function designed to pass through a frequency band centered on, for example, 900 MHz. The A / D converter 104 uses the sampling frequency f s (for example, 3 · 6 G Η z in this embodiment) to convert the RF signal formed by the noise spectrum into digital data. The A / D converter 104 includes a single comparator. Alternatively, the A / D converter 104 includes a three-digit quasi-paper standard applicable to the Chinese National Standard (CNS) A4 specification (2 丨 〇 > < 297 mm) (Please read the precautions on the back before filling in this Page), 1T -Φ-. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs -5- 502494 A7 B7 V. Description of the invention (3) Switching comparator. (Please read the precautions on the back before filling this page) Gate drive circuit 1 06 takes the pulse train from A / D converter 104 and generates the gate drive for the power output stage of the amplifier 100 Of each of the FETs 108 and 10. The output power stage shown includes three inductors LI, L2, and L3, and capacitor C1. When the counterparts in FETs 108 and 110 are closed, this configuration generates separate resonances at nodes A and B, respectively. Through the feedback path 1 12, continuous time feedback is provided to the frequency selection network 1 2. The output signal of the power stage will be sent to the matching network 1 1 4 and the matching network 1 1 4 will send the output RF signal to the antenna 1 1 6 for transmission. The top and bottom F E T 108 and 110 are driven alternately by the gate signals generated by the gate drive circuit 106 at the output of the A / D converter 104. If the FETs 108 and 110 are driven by switching signals such as square waves, the parasitic gate capacitance of each device will cause serious and unnecessary loss. This loss is related to the frequency of the gate signal, the gate capacitance, and the economy. Ministry of Intellectual Property Bureau employee consumer cooperative printing and input voltage squared is proportional to f cv2 loss. Therefore, the inductor 1 1 8 will be connected in series with each F E T gate to resonate to eliminate f c v 2 loss. As described above, the frequency selection network 10 2 cooperates with the continuous time feedback path 1 12 to provide noise spectrum shaping to the amplifier. The network 10 2 includes one or more resonator stages. Although this will reduce these unnecessary losses, there are other undesirable consequences. First, an undesired and substantially overshoot (or down) voltage is generated at the input of each F E T during the gate signal transition. Second, each paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) -6-502494 A7 B7 V. Description of the invention (4) The transition of a gate at each gate is virtually infinite Ground continuous resonance. If the amplifier is only used to generate a regular pattern of 0 and 1 interleaving (frequency and phase modulation applications), this resonance is not a problem because it will cause the required bit pattern (if overshoot / downshift is acceptable). For applications that require aperiodic bit patterns, this resonance interferes with the generation of the required bit patterns. In addition, the use of a fixed timepiece to produce a timepiece for A / D converter 104 (ie, f s) has disadvantages. First, the timing of the resonances at nodes A and B usually needs to match the timing of this time to maximize efficiency. However, in fact, for example, the resonances at nodes A and B tend to vary in frequency due to the response from the matching network 114 and the processing mutation. Second, due to design complexity, it is generally undesirable to have separate timepieces for the A / D converter in this design. As is known, the amplifier configuration of FIG. 1 allows a binary state. With the two-quantization state, there will be a larger number of signal transitions that lead to high drive losses. According to pulse width modulation (pw M) and other modulation techniques, the delay caused by switching and output filter stages must be effectively handled To mitigate the adverse effects of this delay on circuit stability. This is especially true for modulators with fairly high power switching stages, as the delay becomes large relative to the pulse repetition frequency of the loop. The conventional solution to this delay processing will be explained with reference to Figs. 2 and 3. Figure 2 is a block diagram of a typical modulator circuit 200. The logic output of the modulator 2 0 2 will drive the inverting power stage 2 0 4 and the inverting power stage. The paper size applies the Chinese National Standard (CNS) Α4 specification (210 > < 297 mm) (please read the Please note this page and fill in this page) Order Φ .. Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs's Consumer Cooperatives 502494 A7 B7 V. Invention Description (5) (Please read the notes on the back before filling this page) 2 0 4 output meeting Filtered by output filter 2 0 6. The feedback resistor 2 0 8 and the attenuation resistor 2 1 2 are used to introduce negative feedback from the loop output of the modulator 2 0 2. FIG. 3 shows waveforms 3 2 and 3 4 of the second regulator circuit without delay compensation from FIG. 2. As shown, the positive swing of waveform 3 0 (ie, output filter 2 0 6) and the positive swing of waveform 3 0 2 (ie, modulator 202) due to the delay of switching and filter stage introduction. Logic output) is almost in phase. Therefore, in cases where the loop's original design considered negative feedback, the delay would convert it to positive feedback, thus causing the loop to be unstable (not shown). Therefore, a filter capacitor 2 10 is provided in parallel with the feedback resistor 208 for delay compensation. Capacitor 2 10 produces zero in the feedback loop. For high frequencies, the attenuation caused by resistors 208 and 2 1 2 can be bypassed. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. Unfortunately, since this type of delay compensation is performed after the delay introduction of the loop, it is difficult to correct all negative delay results related to loop stability. In fact, this type of compensation technology has only limited results. As a result, the delay tolerance and overall performance of today's modulators have been limited. One solution is to use a feedback signal directly from the logic output of the modulator (that is, before the delay is introduced) combined with the output of the power stage and / or output filter. Unfortunately, although the output of the modulator (302) has similar characteristics to the filter output (304), due to the non-linear nature of the power stage, useful content of the frequency range of the modulator loop There will be significant differences. This will make it difficult to achieve high-fidelity operation when giving back these signals. Based on the above, there is a need for a band-pass amplifier that overcomes the different problems described above. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -8- 502494 A7 B7 V. Description of the invention (6) Invention description (please read the precautions on the back and fill in this purchase first) In one aspect, a band-pass amplifier with a frequency-selective network in the feedback loop is provided. The analog-to-digital converter is coupled to the frequency-selective network. The switching stage is coupled to the analog-to-digital converter to generate a continuous-time output signal. The switching stage includes At least one resonance circuit which resonates at a resonance frequency and generates at least a part of a continuous-time output signal. The continuous-time feedback path will sense and feedback the continuous-time output signal to a frequency selection network. It is also provided for a band-pass amplifier. Switching stage. The switching stage includes first and second switching switches in a half-bridge configuration. Both the first and second switching switches have parasitic capacitances associated with them. The switching stage includes multiple inductors. The second resonant circuit consists of parasitic capacitances. And inductor formation, the two resonance circuits are staggered resonance at the resonance frequency. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs In another aspect, since these inductors need to reduce the fcv 2 loss at the FET gate, the present invention provides a gate drive circuit to control the gate signal technology to overcome the problems of overshoot and resonance. That is, the gate drive The leading and trailing edges of the gate pulses generated by the circuit will be temporarily brought to the logic level, brought back to the original logic level, and then brought back to the desired level when the signal level of the FET gate is settled. To the required logic level. In general, the present invention uses control logic and inductors to charge the logic state of the gate capacitor. Therefore, the present invention provides a switch for switching a switch from a first logic state to a second logic state. Method and device for the state. The switch has a gate terminal, and the gate terminal has a gate capacitor and a resistance associated therewith. According to different embodiments, the Chinese paper standard (CNS) A4 applies to this paper scale. (210X297 mm) -9 -502494 A7 __B7_ V. Description of the invention (7) (Please read the precautions on the back before filling out this page). This resistor can be the input resistance of the switch in series and parallel with the outside. Any combination of resistors. The gate terminal also has a series inductance coupled to it. When the switch is in the first logic state, a pulse corresponding to the second logic is applied to the inductor. The pulse has a first level associated with it Energy, which is high enough to overcome the attenuation caused by the resistance (thus allowing the gate terminal to reach the signal level corresponding to the second logic state, thereby switching the switch to the second logic state), and low enough to slow down the inductance due to inductance And the gate capacitor oscillation (so that the gate terminal is settled at the signal level before the switching switch then transitions to the first logic state). According to another aspect of the invention, one of the output switching stages of the amplifier design is used or More resonance to generate the A / D converter timing signal for the amplifier. According to an embodiment, in a case where the output stage includes two switching devices, the resonance nodes related to the two devices are used in a staggered manner to generate a timepiece message.
Prfe 疏。 經濟部智慧財產局員工消費合作社印製 因此,本發明提供包含取樣電路及至少一切換裝置之 電子裝置。每一切換裝置具有與其輸出端有關的共振電路 。共振電路及至少一切換裝置具有至少一與其相關的共振 振盪。電子裝置又包括時計產生電路,其會以至少一共振 振盪之至少一部份,產生用於取樣電路之時計訊號。 根據本發明的另一態樣,提供放大器結構,其中可致 動多位準切換。亦即,此處所述之放大器結構呈現多於二 之量化狀態。此點係部份藉由並聯訊號路徑以達成,每一 並聯訊號路徑具有其自己的取樣電路。多量化狀態包含無 訊號乾出之狀態,藉以避免上述不需要的切換損失。根據 I紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -10- 502494 A7 B7 五、發明説明(8 ) 特定實施例,用於不同取樣電路之時計訊號會獨立地發展 ,造成其它多種優點。 (請先閲讀背面之注意事項再填寫本頁) 因此,本發明提供電子裝置,其包含至少二取樣電路 ,及並聯之至少二切換級。每一切換級耦合至取樣電路之 一。取樣電路及切換級使得電子裝置能夠呈現多於二之量 化狀態。電子裝置又包含時計產生電路,用以產生用於每 一取樣電路之獨立的時計訊號。 根據本發明的另一觀點,揭示克服上述困難之用於調 變器迴路的回饋技術。此處所述的回饋技術使用低電壓調 變器的輸出以補償後續的功率及濾波器級造成之延遲,同 時,縱使如上所述般訊號內容有差異,仍能取得高水準之 逼真度。藉由將調變器級的輸出濾波,以致於其在有用的 調變器迴路頻率範圍之外的頻率成份會被傳送至回饋路徑 ,而其在有用的迴路範圍之內的頻率成份會被衰減,則發 明可取得此結果。因此,迴路的穩定度會因調變器輸出的 某些部份之回饋而增進,而迴路的輸出頻率之逼真度不會 受此範圍內不需要的調變器輸出成份之不利影響。 經濟部智慧財產局員工消費合作社印製 本發明的回饋濾波器衰減及傳送之頻率會視調變器的 型式而變。舉例而言,對於基頻帶調變器而言,諸如高通 帶濾波器等濾波器會將基頻帶中的頻率衰減並傳送較高的 頻率。對於帶通調變器而言,回饋濾波器可能表現成如同 陷波濾波器,將相關頻帶內的頻率衰減。同樣地,對於帶 除調變器而言,濾波器表現如同帶通濾波器,而對於高通 調變器而言,濾波器表現如同低通濾波器。 11 - 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) 502494 A7 B7 五、發明説明(9 ) (請先閱讀背面之注意事項再填寫本頁) 在容許2 5 0 n s延遲而無本發明改進之調變器迴路 中,加入此處所述之回饋技術,顯示出延遲容許度增加至 大於5 0 0 n s。此意指根據本發明,即使很大的功率裝 置具有擾人的延遲,這些功率裝置仍可用於低電壓調變器 。舉例而言,本發明的特定實施例能夠以非常高的逼真度 將大於1 000W (>lhp !)之功率遞送至4〇hm 的負載。對於1 〇 hm的負載而言(舉例而言,工業用監 視器),這相當於大於5 h p !相較於目前可用之高逼真 調變器迴路,這是相當顯著的改進,且足以用於很多相當 高功率的工業應用。 因此,本發明提供調變器迴路,其具有相關的帶通頻 率範圍並包含具有相關的第一延遲之切換級。調變器迴路 也包含具有回饋輸入之調變器級。調變器級的輸出耦合至 切換級的輸入。第一回饋路徑會耦合於切換級的輸出與調 變器級之間。對應於帶通頻率範圍之陷波濾波器會耦合於 調變器級的輸出與調變器級的回饋輸入之間,以補償第一 延遲。 經濟部智慧財產局員工消費合作社印製 參考說明書的其餘部份及圖式,可進一步瞭解本發明 的本質及優點。 圖式簡述 圖1係雜訊頻譜成型放大器的簡圖。 圖2係調變器迴路的簡化方塊圖。 圖3係顯示來自圖2的調變器迴路之二波形。 本紙張尺度適用中國國家標準(CNS ) A4規格(210 X ;297公釐) -12- 502494 經濟部智慧財產局員工消費合作社印製 附件一(A) ·第90105026號含利申請案令文說明書修正頁 A7 5 月呈 ——-』充 圖4係根據本發明的特定實施例設計之蜂巢式電話的 方塊圖。 圖5係根據本發明的特定實施例設計之數位r f功率 放大器的電路圖。 圖6係根據本發明的另一實施例設計之數位R JP功率 放大器的示意圖。 圖7包含圖7A,7B及7C,係根據本發明的特定 實施例設計之R F放大器的輸出級之更詳細的圖形。 圖8 a係波形序列,說明本發明的特定實施例。 圖8 b係用於本發明之閘極驅動電路的一可能實施之 簡化圖。 圖9 a係本發明的另一特定實施例之簡化圖。 圖9 b係時序圖,說明圖9 a的實施例之操作。 圖9 c係用於圖9 a的實施例之閘極驅動電路的一可 能實施之簡化圖。 圖1 0係顯示根據本發明的特定實施例設計之時計產 生電路。 圖1 1係顯示第二雜訊頻譜成型放大器。 圖1 2係根據發明的特定實施例設計之調變器迴路的 方塊圖。 圖1 3係根據發明的更特定實施例設計之調變器迴路 的方塊圖。 圖1 4係根據發明的又更特定實施例設計之調變器迴 路的方塊圖。 本纸張尺度適用中國國家標準(CNS ) A4規格(210X297公楚) 五、發明说明(1〇) B7 月 (請先閱讀背面之注意事項再填寫本頁) -13- 502494 A7 B7 五、發明説明(11) 圖1 5係根據發明的又更加特定實施例設計之調變器 迴路的方塊圖。 (請先閱讀背面之注意事項再填寫本頁) 圖1 6係根據發明設計的調變器迴路之一般實施例的 方塊圖。 圖1 7係根據發明的特定實施例設計之帶通放大器的 方塊圖。 主要元件對照表 1 0 0 射頻 帶 通 雜 訊頻譜整型放大器 1 0 2 選頻 網 路 1 0 4 A / D 轉 換 器 1 0 6 閘極 驅 動 電 路 1 0 8 場效 電 晶 體 1 1 0 場效 電 晶 體 1 1 2 回饋 路 徑 1 1 4 匹配 網 路 1 1 6 天線 1 1 8 電感 器 2 0 0 調變 器 迴 路 2 0 2 調變 器 2 0 4 反相 功 率 級 2 0 6 輸出 濾 波 器 2 0 8 回饋 電 阻 器 2 10 電容器 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -14- 502494 A7 B7 五、發明説明(12) 2 1 2 衰 減 電 阻 器 4 1 0 蜂 巢 式 電 話 4 1 2 射 頻 4 1 4 調 變 及 解 調 變 區 4 1 6 控 制 器 1¾ 4 1 8 雙 多 工 器 4 2 0 天 線 4 3 0 I F 至 R F 混 its 頻 器 4 3 2 匹 配 網 路 4 3 4 功 率 放大 器 4 3 6 匹 配 網 路 4 4 〇 匹 配 網 路 4 4 2 L N A 放 大 器 4 4 4 匹 配 網 路 4 4 6 R F 至 I F 混 ifcs 頻 器 5 5 〇 C巳巳 頻 網 路 (請先閲讀背面之注意事項再填寫本頁) 訂 經濟部智慧財產局員工消費合作社印製 552 類比對數位轉換器 5 5 4 輸出級 556 回饋迴路 558 數位訊號 600 射頻帶通雜訊頻譜整型放大器 602 選頻網路 604a A/D轉換器 604b A/D轉換器 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -15- 502494 A7 B7 五、發明説明(13) 6 0 6 a 閘極驅動電路 606b 閘極驅動電路 (請先閲讀背面之注意事項再填寫本頁) 608a 場效電晶體 608b 場效電晶體 610a 場效電晶體 610b 場效電晶體 611 共振器電路 611a 共振器電路 6 12 回饋路徑 6 1 3 加法器 614 匹配網路 6 16 天線 902 浮動切換開關 9 0 4 切換開關 1002 比較器 1004 比較器 1006 環形振盪器 1008 多工器 經濟部智慧財產局員工消費合作社印製 1100 射頻放大器 1102 選頻網路 1104a A/D轉換器 1104b A/D轉換器 1106a 閘極驅動電路 1106b 閘極驅動電路 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -16- 502494 經濟部智慧財產局員工消費合作社印製 A7 B7五、發明説明(14) 1108a 場效電晶體 1108b 場效電晶體 1110a 場效電晶體 1110b 場效電晶體 1111 共振器電路 1111a 共振器電路 1112 回饋路徑 1113 加法器 1114 匹配網路 1116 天線 1200 調變器迴路 1202 調變器級 1204 切換級 1206 輸出濾波器 1208 回饋濾波器 1210 總合接頭 1300 基頻頻帶調變器迴路 1302 調變器 1304 切換級 1306 輸出濾波器 1308 濾波器電容器 1310 串連電阻器 1312 電阻器 1314 電阻器 (請先閱讀背面之注意事項再填寫本頁) 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -17- 502494 A7 B7 五、發明説明(15) 經濟部智慧財產局員工消費合作社印製 1 4 〇 0 基 頻 頻 帶 調 變器迴路 1 4 〇 2 調 變 器 1 4 〇 4 切 換 級 1 4 〇 6 輸 出 濾 波 器 1 4 0 8 濾 波 器 電 容 器 1 4 1 0 串 連 電 阻 器 1 4 1 2 電 阻 器 1 4 1 4 電 阻 器 1 4 1 6 電 容 器 1 5 0 〇 調 變 器 迴 路 1 5 0 2 S田 m 變 器 1 5 〇 4 切 換 級 1 5 〇 6 輸 出 濾 波 器 1 5 〇 8 濾 波 器 電 容 器 1 5 1 0 串 連 電 阻 器 1 5 1 2 電 阻 器 1 5 1 4 電 阻 器 1 5 1 6 電 容 器 1 5 2 0 電 阻 器 1 5 2 2 電 容 器 1 6 〇 〇 三田 δ周 變 器 迴 路 1 6 0 2 調 變 器 1 6 〇 4 切 換 級 1 6 0 6 輸 出 濾 波 器 級 (請先閲讀背面之注意事項再填寫本頁) 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -18- 502494 A7 B7 五、發明説明(16) 1 6 0 8 4埠網路 1610 增益控制 (請先閲讀背面之注意事項再填寫本頁) 1700 射頻帶通雜訊頻譜整型放大器 1702 選頻網路 1704 A/D轉換器 1706 閘極驅動電路 1708 場效電晶體 1710 場效電晶體 17 12 回饋路徑 1714 匹配網路 1716 天線 1718 回饋路徑 1720 陷波濾波器 特定實施例詳述 經濟部智慧財產局8工消費合作社印製 參考圖4,其顯示根據本發明設計之蜂巢式電話 4 1 〇的簡化方塊圖。蜂巢式電話4 1 0包含RF區 412、調變及解調變區414、包含DSP、RAM、 使用者介面、用以產生基頻帶訊號之基頻帶電路及其它用 於蜂巢式電話中的標準電路之控制器區4 1 6、雙工器或 T / R切換開關4 1 8 (雙工器切換開關用於同時發送及 接收之蜂巢式電話,T / R切換開關係用於交錯地發送及 接收之蜂巢式電話)、以及天線4 2 〇。在發送期間,控 制器416中的基頻帶電路負責產生基頻帶訊號,基頻帶 I紙張尺度適用巾關家縣(CNS) A4規格(21Gx297公襲) -19 - 502494 A7 B7 五、發明説明(17) (請先閲讀背面之注意事項再填寫本頁) 訊號典型上是代表要發送的資訊之位元串。然後,於盒 4 1 4中以中頻調變訊號調變基頻帶訊號,然後提供給 R F區4 1 2。由於調變及解調變區4 1 4、控制器 416、雙工器/T/R切換開關418及天線420之 功能在此技藝中係習知的,所以,此處將不詳述。 RF區412包含IF至RF混頻器430、匹配網 路(MN) 4 3 2、用以放大RF訊號之功率放大器( PA) 434、及另一匹配網路(MN) 436,混頻器 4 3 0係用以藉由經由調變器4 1 4自控制器4 1 6接收 之經過調變的訊號,產生R F訊號(舉例而言,9 0 0 0 MHz)。在發送期間,放大器434會大RF訊號並將 其經由匹配網路4 3 6及雙工器/T/R切換開關4 1 8 提供給天線4 2 0。在接收側上,R F區4 1 2包含匹配 網路440、LNA放大器442、另一匹配網路444 、r F至I F混頻器4 4 6,匹配網路4 4 0係用以接收 天線4 2 0收到的R F訊號(舉例而言,9 8 0 Μ Η z ) 經濟部智慧財產局員工消費合作社印製 ,R F至I F混頻器會將R F訊號混合成I F範圍,然後 將其提供給區塊4 1 4中的解調變器。解調變器4 1 4會 將發送的基頻帶資訊解調變並將其提供給控制器4 1 6。 由於混頻器電路430和446、LNA 442、及匹 配網路432、 440和444均屬習知,所以此處不再 詳述。 參考圖5,其顯示根據本發明的特定實施例設計之數 位R F功率放大器4 3 4。根據不同的特定實施例,可依 本紙張尺度適用中國國家標準(CNS ) A4規格(21〇X:297公釐) -20 - 502494 A7 B7 五、發明説明(18) 據1 9 9 8年七月七號頒予之美國專利號5,7 7 7, 5 1 2 「METHOD AND APPARATUS FOR OVERSAMPLED, NOISE-SHAPING, MIXED-SIGNAL PROCESSING」中揭示的 技術,設計放大器4 3 4。放大器4 3 4包含在所需發送 頻率具有高Q値之選頻網路5 5 0、類比對數位轉換器( A/D) 552、輸出級554、及回饋迴路556, 回饋迴路5 5 6係提供從輸出級5 5 4至選頻網路5 5 0 之連續時間回饋訊號。A/D轉換器5 5 2會取樣選頻網 路5 5 0之輸出並產生提供給輸出級5 5 4之數位訊號 558。在一實施例中,A/D轉換器552係比較器, 其以3 · 6GHz的頻率取樣,以應用至需用900 Μ Η z的發送頻率之應用。在其它可能實施例中,對於 900MHz訊號的發送而言,取樣頻率(f s)的範圍 從最小1 · 8GHz至3 · 6GHz或更高。根據一此種 可能實施例,A / D轉換器5 5 2係與延遲線串聯之未計 時的比較器(亦即,i s 〜,基本上操作類似史米特 觸發器。一般而言,取樣頻率(f s )較佳地至少爲所需 發送頻率的2倍。 輸出級5 5 4的一可能實施係包含二電晶體T 1和 丁2、電感器LI、 L2和L3、電容器Cl、 C2和 C 3、及前置驅動器D。前置驅動器D係用以緩衝訊號 5 5 8及分別提供訊號5 5 8和其互補給電晶體T 1和電 晶體T 2之閘極。T 1的汲極耦合至V c c ,而源極耦合 至節點A。電容器C 1典型上是位於電晶體T 1的源極與 本紙張又度適用中國國家標準(CNS ) A4規格(210X297公釐) (請先閲讀背面之注意事項再填寫本頁)Prfe sparse. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economics Therefore, the present invention provides an electronic device including a sampling circuit and at least one switching device. Each switching device has a resonant circuit associated with its output. The resonance circuit and at least one switching device have at least one resonance oscillation associated therewith. The electronic device further includes a timepiece generating circuit which generates at least a portion of at least one resonance oscillation to generate a timepiece signal for the sampling circuit. According to another aspect of the present invention, there is provided an amplifier structure in which multi-level switching can be activated. That is, the amplifier structure described herein exhibits more than two quantized states. This is achieved in part through parallel signal paths, each of which has its own sampling circuit. The multi-quantization state includes a state where no signal is output, so as to avoid the unnecessary switching loss mentioned above. Applicable to China National Standard (CNS) A4 specification (210X297 mm) according to I paper size -10- 502494 A7 B7 V. Description of the invention (8) In specific embodiments, the timepiece signal for different sampling circuits will develop independently, causing other Multiple advantages. (Please read the precautions on the back before filling this page) Therefore, the present invention provides an electronic device including at least two sampling circuits and at least two switching stages connected in parallel. Each switching stage is coupled to one of the sampling circuits. The sampling circuit and switching stage enable the electronic device to exhibit more than two quantization states. The electronic device further includes a timepiece generating circuit for generating an independent timepiece signal for each sampling circuit. According to another aspect of the present invention, a feedback technique for a regulator circuit that overcomes the above-mentioned difficulties is disclosed. The feedback technology described here uses the output of the low-voltage modulator to compensate for the delay caused by subsequent power and filter stages. At the same time, even if the signal content is different as described above, a high level of fidelity can still be obtained. By filtering the output of the modulator stage so that its frequency components outside the useful loop frequency range are transmitted to the feedback path, and its frequency components within the useful loop range are attenuated , The invention can achieve this result. Therefore, the stability of the loop will be improved by feedback from some parts of the modulator output, and the fidelity of the output frequency of the loop will not be adversely affected by the unwanted output components of the modulator in this range. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs The attenuation and transmission frequency of the feedback filter of the present invention will vary depending on the type of modulator. For example, for baseband modulators, filters such as high-pass filters attenuate frequencies in the baseband and transmit higher frequencies. For a band-pass modulator, the feedback filter may behave like a notch filter, attenuating frequencies in the relevant frequency band. Similarly, the filter behaves like a band-pass filter to a band-elimination modulator, and to a low-pass filter to a high-pass modulator. 11-This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) 502494 A7 B7 V. Description of the invention (9) (Please read the precautions on the back before filling this page) The delay of 2 50 0 ns is allowed Adding the feedback technology described herein to the modulator circuit without the improvement of the present invention shows that the delay tolerance is increased to more than 500 ns. This means that according to the present invention, even if very large power devices have disturbing delays, these power devices can still be used in low voltage modulators. For example, certain embodiments of the present invention are capable of delivering power greater than 1,000 W (> lhp!) To a load of 40 hm with very high fidelity. For a load of 10hm (for example, an industrial monitor), this is equivalent to more than 5 hp! This is a significant improvement over the currently available high-fidelity modulator circuits and is sufficient for Many fairly high power industrial applications. Accordingly, the present invention provides a modulator circuit having an associated band-pass frequency range and including a switching stage having an associated first delay. The modulator circuit also includes a modulator stage with a feedback input. The output of the modulator stage is coupled to the input of the switching stage. The first feedback path is coupled between the output of the switching stage and the modulator stage. A notch filter corresponding to the band-pass frequency range is coupled between the output of the modulator stage and the feedback input of the modulator stage to compensate for the first delay. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. With reference to the rest of the manual and the drawings, you can further understand the nature and advantages of the present invention. Brief description of the figure Figure 1 is a simplified diagram of a noise spectrum shaping amplifier. Figure 2 is a simplified block diagram of the modulator circuit. FIG. 3 shows the second waveform of the modulator circuit from FIG. 2. This paper size applies to Chinese National Standard (CNS) A4 (210 X; 297 mm) -12- 502494 Printed by Annex I (A) of the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, No. 90105026, Profit Application Application Order Description Revised page A7 was presented in May ----- "Fill in Figure 4 is a block diagram of a cellular phone designed according to a specific embodiment of the present invention. FIG. 5 is a circuit diagram of a digital r f power amplifier designed according to a specific embodiment of the present invention. FIG. 6 is a schematic diagram of a digital R JP power amplifier designed according to another embodiment of the present invention. Fig. 7 contains Figs. 7A, 7B, and 7C, which are more detailed diagrams of the output stages of an RF amplifier designed according to a particular embodiment of the present invention. Figure 8a is a waveform sequence illustrating a specific embodiment of the invention. Fig. 8b is a simplified diagram of a possible implementation of a gate driving circuit for the present invention. Fig. 9a is a simplified diagram of another specific embodiment of the present invention. Fig. 9b is a timing diagram illustrating the operation of the embodiment of Fig. 9a. Fig. 9c is a simplified diagram of a possible implementation of the gate driving circuit for the embodiment of Fig. 9a. Fig. 10 shows a clock generation circuit designed according to a specific embodiment of the present invention. Figure 11 shows the second noise spectrum shaping amplifier. Figure 12 is a block diagram of a modulator circuit designed in accordance with a particular embodiment of the invention. Figure 13 is a block diagram of a modulator circuit designed in accordance with a more specific embodiment of the invention. Figure 14 is a block diagram of a modulator circuit designed in accordance with yet a more specific embodiment of the invention. This paper size applies to China National Standard (CNS) A4 specification (210X297). 5. Description of invention (1〇) BJuly (please read the precautions on the back before filling this page) -13- 502494 A7 B7 5. Invention Explanation (11) FIG. 15 is a block diagram of a modulator circuit designed according to yet a more specific embodiment of the invention. (Please read the notes on the back before filling out this page) Figure 16 is a block diagram of a general embodiment of a modulator circuit designed according to the invention. Figure 17 is a block diagram of a bandpass amplifier designed in accordance with a particular embodiment of the invention. Comparison table of main components 1 0 0 RF bandpass noise spectrum shaping amplifier 1 0 2 Frequency selection network 1 0 4 A / D converter 1 0 6 Gate driver circuit 1 0 8 Field effect transistor 1 1 0 Field effect Transistor 1 1 2 Feedback path 1 1 4 Matching network 1 1 6 Antenna 1 1 8 Inductor 2 0 0 Modulator loop 2 0 2 Modulator 2 0 4 Inverting power stage 2 0 6 Output filter 2 0 8 Feedback resistor 2 10 Capacitor printed by the Intellectual Property Bureau of the Ministry of Economic Affairs Consumer Cooperatives This paper is printed in accordance with China National Standard (CNS) A4 (210X297 mm) -14- 502494 A7 B7 V. Description of the invention (12) 2 1 2 Attenuation resistor 4 1 0 Cellular phone 4 1 2 Radio frequency 4 1 4 Modulation and demodulation zone 4 1 6 Controller 1 ¾ 4 1 8 Dual multiplexer 4 2 0 Antenna 4 3 0 IF to RF mixer its frequency converter 4 3 2 Matching network 4 3 4 Power amplifier 4 3 6 Matching network 4 4 〇 Matching network 4 4 2 LNA amplifier 4 4 4 Matching network 4 4 6 RF to IF mixed ifcs frequency converter 5 5 〇C frequency network (Please read the precautions on the back before filling out this page) Order the 552 analog to digital converter printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 5 5 4 output stage 556 feedback loop 558 digital signal 600 RF bandpass noise spectrum shaping amplifier 602 frequency selection network 604a A / D converter 604b A / D converter This paper size applies to China National Standard (CNS) A4 specification (210X297 (Mm) -15- 502494 A7 B7 V. Description of the invention (13) 6 0 6 a Gate driver circuit 606b Gate driver circuit (please read the precautions on the back before filling this page) 608a Field effect transistor 608b Field effect Transistor 610a Field Effect Transistor 610b Field Effect Transistor 611 Resonator Circuit 611a Resonator Circuit 6 12 Feedback Path 6 1 3 Adder 614 Matching Network 6 16 Antenna 902 Floating Switch 9 0 4 Switch 1002 Comparator 1004 Compare 1006 Ring oscillator 1008 Multiplexer Printed by the Intellectual Property Bureau of the Ministry of Economic Affairs Consumer Cooperative 1100 RF amplifier 1102 Frequency selection network 1104a A / D converter 1104b A / D converter 1106a Pole drive circuit 1106b Gate drive circuit This paper size is applicable to Chinese National Standard (CNS) A4 specification (210X297 mm) -16- 502494 Printed by A7 B7, Consumer Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs 5. Description of invention (14) Effect transistor 1108b Field effect transistor 1110a Field effect transistor 1110b Field effect transistor 1111 Resonator circuit 1111a Resonator circuit 1112 Feedback path 1113 Adder 1114 Matching network 1116 Antenna 1200 Modulator circuit 1202 Modulator stage 1204 Switch Stage 1206 output filter 1208 feedback filter 1210 total joint 1300 baseband frequency band modulator circuit 1302 modulator 1304 switching stage 1306 output filter 1308 filter capacitor 1310 series resistor 1312 resistor 1314 resistor (please first Read the notes on the reverse side and fill in this page) This paper size is applicable to Chinese National Standard (CNS) A4 specification (210X297 mm) -17- 502494 A7 B7 V. Description of invention (15) Printed by the Consumer Cooperative of Intellectual Property Bureau, Ministry of Economic Affairs 1 4 〇 0 fundamental frequency band modulator circuit 1 4 〇 2 modulator 1 4 〇 4 switching stage 1 4 〇6 Output filter 1 4 0 8 Filter capacitor 1 4 1 0 Series resistor 1 4 1 2 Resistor 1 4 1 4 Resistor 1 4 1 6 Capacitor 1 5 0 〇 Modulator circuit 1 5 0 2 S Tianm transformer 1 5 〇4 switching stage 1 5 〇6 output filter 1 5 〇8 filter capacitor 1 5 1 0 series resistor 1 5 1 2 resistor 1 5 1 4 resistor 1 5 1 6 capacitor 1 5 2 0 Resistor 1 5 2 2 Capacitor 1 6 〇〇 Mita delta cycle converter circuit 1 6 0 2 Modulator 1 6 〇 4 Switching stage 1 6 0 6 Output filter stage (Please read the precautions on the back first (Fill in this page again) This paper size applies Chinese National Standard (CNS) A4 specification (210X297 mm) -18- 502494 A7 B7 V. Description of the invention (16) 1 6 0 8 4-port network 1610 gain control (please read first Note on the back, please fill out this page again) 1700 RF Band Pass Noise Spectrum Shaping Amplifier 1702 Frequency Select Network 1704 A / D Converter 1706 Gate Driver Circuit 1708 Field Effect Transistor 1710 Field Effect Transistor 17 1 2 Feedback path 1714 Matching network 1716 Antenna 1718 Feedback path 1720 The specific embodiment of the notch filter is described in detail with reference to Figure 4 printed by the Intellectual Property Bureau of the Ministry of Economic Affairs 8 Industrial Cooperative Cooperative, which shows a cellular phone 4 1 designed according to the present invention. Simplified block diagram. Cellular phone 4 1 0 includes RF area 412, modulation and demodulation area 414, including DSP, RAM, user interface, baseband circuits to generate baseband signals, and other standard circuits used in cellular phones Controller area 4 1 6 、 Duplexer or T / R switch 4 1 8 (The duplexer switch is used for the cellular phone that sends and receives at the same time, and the T / R switch is used for the interleaved sending and receiving. Cell phone), and antenna 4 2 0. During transmission, the baseband circuit in the controller 416 is responsible for generating the baseband signal. The baseband I paper size is suitable for Guanjia County (CNS) A4 specification (21Gx297 public attack) -19-502494 A7 B7 V. Description of the invention (17 ) (Please read the notes on the back before filling out this page) The signal is typically a bit string representing the information to be sent. Then, the baseband signal is modulated with an IF modulation signal in the box 4 1 4 and then provided to the R F area 4 1 2. Since the functions of the modulation and demodulation area 4 1 4, the controller 416, the duplexer / T / R switch 418 and the antenna 420 are well known in this technology, they will not be described in detail here. RF area 412 contains IF to RF mixer 430, matching network (MN) 4 3 2. Power amplifier (PA) 434 to amplify RF signals, and another matching network (MN) 436, mixer 4 3 0 is used to generate an RF signal by using a modulated signal received from the controller 4 1 6 via the modulator 4 1 4 (for example, 9 0 0 MHz). During transmission, the amplifier 434 will generate an RF signal and provide it to the antenna 4 2 0 via the matching network 4 3 6 and the duplexer / T / R switch 4 1 8. On the receiving side, RF zone 4 1 2 contains matching network 440, LNA amplifier 442, another matching network 444, r F to IF mixer 4 4 6 and matching network 4 4 0 is used to receive antenna 4 RF signal received at 20 (for example, 980 Μ Η z) Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, the RF to IF mixer will mix the RF signal into the IF range, and then provide it to Demodulator in block 4 1 4. The demodulator 4 1 4 will demodulate the transmitted baseband information and provide it to the controller 4 1 6. Since mixer circuits 430 and 446, LNA 442, and matching networks 432, 440, and 444 are all known, they will not be described in detail here. Referring to Fig. 5, there is shown a digital RF power amplifier 4 3 4 designed in accordance with a particular embodiment of the present invention. According to different specific embodiments, the Chinese National Standard (CNS) A4 specification (21 ×: 297 mm) can be applied according to the paper size. -20-502494 A7 B7 V. Description of the invention (18) According to 1989 7 The technology disclosed in US Patent No. 5,7 7 7, 5 1 2 issued on the 7th of July, "METHOD AND APPARATUS FOR OVERSAMPLED, NOISE-SHAPING, MIXED-SIGNAL PROCESSING", designed the amplifier 4 3 4. The amplifier 4 3 4 includes a frequency selective network 5 5 0 with a high Q 値 at the required transmission frequency, an analog to digital converter (A / D) 552, an output stage 554, and a feedback circuit 556, a feedback circuit 5 5 6 series. Provide continuous time feedback signal from output stage 5 54 to frequency selective network 5 50. The A / D converter 5 5 2 samples the output of the frequency selective network 5 5 0 and generates a digital signal 558 that is provided to the output stage 5 5 4. In one embodiment, the A / D converter 552 is a comparator, which samples at a frequency of 3.6 GHz to be applied to an application that requires a transmission frequency of 900 MHz. In other possible embodiments, for the transmission of a 900 MHz signal, the sampling frequency (f s) ranges from a minimum of 1.8 GHz to 3.6 GHz or higher. According to one such possible embodiment, the A / D converter 5 5 2 is an untimed comparator (that is, is ~) connected in series with the delay line, which basically operates like a Schmitt trigger. Generally speaking, the sampling frequency (Fs) is preferably at least twice the required transmission frequency. One possible implementation of the output stage 5 5 4 includes two transistors T 1 and D 2, inductors LI, L 2 and L 3, capacitors Cl, C 2 and C 3. And the pre-driver D. The pre-driver D is used to buffer the signal 5 5 8 and provide the signal 5 5 8 and its complementary power supply transistor T 1 and the gate of the transistor T 2 respectively. The drain of T 1 is coupled to V cc, and the source is coupled to node A. Capacitor C 1 is typically located at the source of transistor T 1 and this paper is again compatible with China National Standard (CNS) A4 (210X297 mm) (Please read the back (Please fill in this page again)
、tT 經濟部智慧財產局員工消費合作社印製 -21 - 502494 Α7 Β7 五、發明説明(19) (請先閱讀背面之注意事項再填寫本頁) 汲極之間的寄生電容器。T 2的汲極耦合至節點B ,而源 極耦合至接地。電容器C 2典型上是電晶體T 2的源極與 汲極之間的寄生電容器。電感器L 1耦合於節點A與節點 B之間,電感器L 2和L 3分別耦合於電容器C 3與節點 A之間。 經濟部智慧財產局員工消費合作社印製 在操作期間,由A/D轉換器552產生的數位訊號 5 5 8會根據正發送之資訊而在高與低之間轉換。由於訊 號5 5 8與其互補會分別提供給電晶體T 1及T 2的閘極 ,所以,一電晶體總是斷開,而另一者會視訊號5 5 8的 狀態而閉接。舉例而言,當訊號· 5 5 8轉變成低時,T 1 閉接而T 2斷開。當此點發生時,源極節點A會因輸出級 5 5 4內的共振電路之形成而共振。此共振電路係由c 1 、三電感器LI、L2、和L3、及經由T2被拉至接地 之節點B所形成。如此實行之驅動器電5 5 4在節點A和 B包含二分別的共振。取決於訊號5 5 8的狀態,一節點 共振,而另一節點會被箝位。在一實施例中,共振會轉變 成以3 · 6 G Η z的取樣頻率共振。藉由適當地選取電感 器LI、L2、和L3、以及電容器Cl、C2和03的 値,可達成此點。根據特定實施例,選取C 3以使R F頻 率通過。 切換級5 5 4的輸出會提供給匹配網路4 3 6 ,匹配 網路4 3 6係作爲在發送頻帶操作之帶通灑波器。由於在 上述實施例中,天線4 2 0以9 0 0 Μ Η ζ發送,所以, 匹配網路4 3 6的「追蹤」功能需要符合此頻率。在一實 本紙張尺度適用中國國家標準(CNS ) Α4規格(210X297公釐) -22- 502494 A7 B7 五、發明説明(20) (請先閱讀背面之注意事項再填寫本頁) 施例中,此點之達成係藉由選取L 1、L 2、及L 3、以 及C3的値,以致於共振電路具有約900MHz的看入 匹配網路4 3 6之轉換函數,以致於T 1和T 2所產生的 輸出位元樣式具有位於9 0 ΟΜΗ z的能量成份。換言之 ,匹配網路4 3 6必須提供足夠的增益以確保位元樣式具 有用於天線的歐姆位準(典型上爲5 0歐姆)之處於 900MHz之充份能量。 在另一實施例,匹配網路4 3 6使用含有功率放大器 4 3 4及其它被動元件之晶片上的接線,以產生匹配網路 以提供最佳功率轉換給天線4 2 0並將天線的歐姆値轉換 成可從給定的電源電壓取得所需功率位準之阻抗。這將需 要具有相當窄的頻寬之相當高Q値濾波器。在又另一實施 例中,功率放大器4 3 4會設計成具有橋接式輸出。在天 線4 2 0具有單一端輸出之應用中,可使用平衡-不平衡 轉換器或被動L C混頻器。 圖6係顯示根據本發明以及上述美國專利號 5, 777, 512中所述的技術設計之另一RF帶通雜 經濟部智慧財產局員工消費合作社印製 訊頻譜整型放大器6 0 0。如同將說明般,R F放大器 6 0 0會配置成實行多位準切換。RF放大器6 0 0包含 選頻網路6 0 2,其使用連續時間回饋,將經過調變的 R F輸入施以雜訊頻譜整型。根據特定實施例,網路 6 0 2包括至少一共振器級,其具有設計成使9 0 0 ΜΗ z爲中心之頻帶通過之轉換函數。 二A/D轉換器6 0 4 a及6 0 4 b會以額定取樣頻 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -23- 502494Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs -21-502494 Α7 Β7 V. Description of the invention (19) (Please read the precautions on the back before filling this page) Parasitic capacitor between the drains. The drain of T 2 is coupled to node B, and the source is coupled to ground. Capacitor C 2 is typically a parasitic capacitor between the source and the drain of transistor T 2. Inductor L 1 is coupled between node A and node B, and inductors L 2 and L 3 are coupled between capacitor C 3 and node A, respectively. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs During operation, the digital signal 5 5 8 generated by the A / D converter 552 will switch between high and low based on the information being sent. Since the signal 5 5 8 and its complement are provided to the gates of the transistors T 1 and T 2 respectively, one transistor is always disconnected and the other is closed depending on the state of the signal 5 5 8. For example, when the signal 5 5 8 goes low, T 1 is closed and T 2 is open. When this occurs, the source node A will resonate due to the formation of a resonance circuit in the output stage 5 54. This resonant circuit is formed by c 1, three inductors LI, L2, and L3, and node B which is pulled to ground via T2. The driver circuit 5 5 4 thus implemented contains two separate resonances at nodes A and B. Depending on the state of the signal 5 5 8, one node resonates and the other node is clamped. In one embodiment, the resonance is transformed into a resonance at a sampling frequency of 3 · 6 G Η z. This can be achieved by appropriately selecting the inductors LI, L2, and L3, and 电容器 of the capacitors Cl, C2, and 03. According to a particular embodiment, C 3 is selected to pass the R F frequency. The output of the switching stage 5 5 4 is provided to the matching network 4 3 6, which is used as a band-pass sprinkler operating in the transmission band. Since in the above embodiment, the antenna 420 is transmitted at 900 MHz, the "tracking" function of the matching network 4 36 needs to conform to this frequency. In the actual paper size, the Chinese National Standard (CNS) A4 specification (210X297 mm) -22-502494 A7 B7 V. Description of the invention (20) (Please read the precautions on the back before filling this page) In the example, This is achieved by selecting L1, L2, and L3, and C3, so that the resonant circuit has a conversion function of approximately 900MHz into the matching network 4 3 6 so that T 1 and T 2 The resulting output bit pattern has an energy component located at 900 MHz. In other words, the matching network 4 3 6 must provide sufficient gain to ensure that the bit pattern has an ohmic level (typically 50 ohms) for the antenna with sufficient energy at 900 MHz. In another embodiment, the matching network 4 3 6 uses wiring on a chip containing a power amplifier 4 3 4 and other passive components to generate a matching network to provide the best power conversion to the antenna 4 2 0 and the antenna's ohms値 is converted into an impedance that can obtain the required power level from a given supply voltage. This will require a fairly high Q 値 filter with a fairly narrow bandwidth. In yet another embodiment, the power amplifier 4 3 4 is designed to have a bridged output. In applications where the antenna 4 2 0 has a single-ended output, a balanced-unbalanced converter or a passive LC mixer can be used. Fig. 6 shows another RF band pass designed by the present invention and the technology described in the above-mentioned U.S. Patent No. 5,777,512, which is printed by a consumer cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs and a printed spectrum shaping amplifier 600. As will be explained, the RF amplifier 600 is configured to perform multi-level switching. The RF amplifier 600 includes a frequency-selective network 602, which uses continuous time feedback to apply a modulated R F input to a noise spectrum shaping. According to a particular embodiment, the network 600 includes at least one resonator stage having a transfer function designed to pass a frequency band centered at 900 MHz. The two A / D converters 6 0 4 a and 6 0 4 b will use the rated sampling frequency. The paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -23- 502494
經濟部智慧財產局員工消費合作社印製 五、發明説明(21) 率f S (亦即f s i及f s 2 ),使用獨立產生的時計訊 號將經過雜訊頻譜整型的RF訊號_成數位資t㈣ 特定實施例,額定取樣頻率f sm3 _ 6eHZ。根據— 貫施例,A / D轉換器6 〇 4 a及6 o 4 b包括用以實施 三位準切換之二比較器。 聞極驅動電路6 0 6 a及6 〇 6 b會分別從A / D轉 換器6 0 4 a及6 〇 4 b取得脈沖列,並產生用於其電晶 體對(亦即FET 608a及6工〇3或1?]2丁 6 0 8 b及6 1 〇 b )之閘極驅動。每一對電晶體具有分 別導因於共振器電路6 1 1和6 1 a之二分別共振。亦 即,包括F E T 6 0 8和6 1 Q之功率級在節點a和B 具有分別的共振,而包括FET 608a和610a之 級在節點A ’和B ’具有分別的共振。 連續時間回饋會經由回饋路徑6 1 2和加法器,提供 給選頻網路6 0 2。功率級的輸出訊號會傳送給匹配網路 6 1 4,匹配網路6 1 4會將輸出R F訊號傳送給天線 6 1 6以用於發送。 具有用於轉換器6 0 4 a和6 0 4 b之二比較器,會 允許數位訊號具有三量化狀態,亦即,三位準切換,而非 二位準切換。舉例而言,根據放大器4 3 4的二位準量化 狀態,可能會有很多造成高驅動損失之訊號過渡。相對地 ,根據三狀態,當無訊號輸出以避免此不需要的切換損失 時,可選取“ 〇 ”狀態。 圖7包括圖7A,7B及7C,係顯示切換級之詳細圖,根據本發明的特定實施 本紙張尺度適用中.國國家標準(CNS ) A4規格(210x297公釐) ---------%II (請先閱讀背面之注意事項再填寫本頁) 訂 •Γ. -24- 502494 A7 B7 經濟部智慧財產局員工消費合作社印製 五、 發明説明( 22 Ο 1 1 例 ,其 可 由圖 6 的 R F 放 大 器 使 用 〇 但 是 y 應 瞭 解 , 在 不 1 1 悖 離發 明 的範 圍 之 下 圖 7 的 切 換 級 以 及 圖 5 和 6 之 放 大 1 I 器 的切 換 級可 用 於 不 同 的放 大 器 型 式 中 〇 亦即 本 發 明 的 ^-V 請 1 1 I 切 換級 配 置適 用 於 R F 範 圍 之 外 的 很 多 應 用 〇 先 閱 讀 1 1 雖 妖 j\w 本發 明 係以 蜂 巢 式 電 話 應 用 作 爲 說 明 但 是 應 背 面 1 瞭 解可 以 以廣 泛 的 內 涵 及 其 它 實 施 例 實 施 本 發 明 〇 舉 例而 注 意 | 言 ,本 發 明的 R F 放 大 器 可 用 於 蜂 巢 式 電 話 基 地 台 中 或 需 Ψ 項 再 1 I 要 有效 率 的R F 放大 器 之任何其 它 應 用 〇 本 發 明 的 數 位放 填 寫 本 • I 大 器也 可 用於 諸 如 1 • 8 G Η Z 或 3 • 0 G Η Ζ 或 較 高 頻 頁 1 1 率 之R F 傳輸 〇 在 這 itb 實 施 例 中 , 藉 由 增 加根 據 上 述 比例 1 1 之 取樣 率 (f S ) 可取 得 較 高 傳 送 速 率 〇 電 晶 體 T 1 及 1 I T 2也 可 爲一 & 不 同 型 式的 裝 置 , 包含 Μ E S F Ε 丁 訂 I Η B T CM 〇 S 、 或 N Μ 〇 S 並 可 由 一 些 不 同 的 製 程 1 1 I 及 材料 ( 包含 G a A S 、 S i G e 或 標 準 矽 ) 製 成 〇 因 1 1 此 ,可 瞭 解此 處 所 提 供 的 實 施 例 僅 爲 舉 例 說 明 , 發 明 之 真 1 | 正 範圍 及 精神 應 由 串 請 專 利 範 圍 決 定 0 # 1 根 據 特定 實 施 例 圖 1 之 雜 訊 頻 譜 整 型 放 大 器 的 閘 極 驅 動電 路 10 6 根 據 本 發 明 的 原 理 操 作 〇 參 考 顯 示 用 於 1 1 F E T 之 一的 閘 極 訊 號 之 上 升及下 降 邊 緣 之 圖 8 a 可 瞭 1 1 解 此點 〇 閘極 訊 號 8 0 2 係 由 閘 極 驅 動 電 路 1 0 6 產 且 訊 1 I Drfe· Wl 8 0 4 位於 F E T 的 閘 極 〇 產 生 所 示 訊 號 8 0 4 以 用 於 1 1 I F E T 10 8 和 1 1 0 中 任 一 者 或 二 者 ’ 其 一 基 本 上 爲 1 1 另 一者 之 互補 0 用 於 閘 極 驅 動 電 路 1 0 6 之 一 可能 實 施 係 1 1 顯 示於 圖 8 b 中 〇 但 是 應 瞭 解 在 不 悖 離 發 明 的 範 圍 之 1 1 1 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -25- 502494 A7 B7 五、發明説明(23) 下,此處所述之本實施例及其它實施例可使用多種驅動電 路。 (請先閲讀背面之注意事項再填寫本頁) 根據一實施例,正好足夠的能量被導入F E T閘極處 的共振中,以致於閘極處的訊號8 0 4會處於所需的輸入 邏輯位準。在閘訊號8 0 2中用於每一邏輯過渡之初始脈 沖的寬度會決定此能量並由閘極驅動電路1 0 6控制。亦 即,閘極驅動電路1 0 6係用以接收來自A / D轉換器 1 0 4 (舉例而言,可爲比較器)之脈沖列8 0 6並產生 閘極訊號,其中,來自A / D轉換器1 〇 4之脈沖列的每 一正及負的過渡對應於圖8 a中所示的對應過渡序列。 所以,對於從訊號8 0 6的邏輯” 0 ”至邏輯” 1 ”之過渡 ,閘極驅動電路1 0 6會在訊號8 0 2中產生造成訊號 8 0 4處於邏輯1之寬度W 1之初始脈沖。然後,在訊號 8 0 6的下一過渡之前·在某一稍晚的點處,將訊號 8 0 2再度帶至高位準(亦即,第三過渡)。在該點處, 發生類似的轉換序列而將過渡帶至邏輯“ 〇 ” 。如同所見 經濟部智慧財產局員工消費合作社印製 ,此舉不僅可防止來自共振之振盪,也能解決超調/下越 問題。 由於有些電阻與每一FET的輸入串聯,所以,訊號 8 0 2中的每一初始過渡脈沖之寬度W需要足夠寬以克服 此電阻導入的衰減因素。否則,訊號8 0 4的位準永遠不 會達到所需的邏輯位準。另一方面,也希望寬度W不會太 寬以免發生顯著的超調或振盪。因此,根據本發明的特定 實施例,可提供適應回饋控制,其會監督訊號8 0 4無法 ^紙張尺度適用中國國家標準(CNS ) A4規格(210X29?公釐) ~ -26 - 502494 A7 B7 五、發明説明(24) 達到或克服所需的邏輯位準之程度,並使用此資訊以經由 閘極驅動電路1 〇 6以調整訊號8 0 2的過渡脈沖之時序 〇 (請先閱讀背面之注意事項再填寫本頁) 根據又一特定實施例,適應迴路包含量測F E T閘極 處的訊號位準是否太高或太低之至少一比較器,及使用 DAC,調整閘極驅動電路1 〇 6中一或更多延遲線路的 偏壓電流以調整訊號8 〇 2的過渡脈沖的時序,藉以校正 過多或不足。 根據不同的實施例,藉由閉接驅動電路,可達到狀態 2a及2b。電感器電流將提供過渡2a、 2b、 3a及 經濟部智慧財產局員工消費合作社印製 3 b。這意指用於訊號8 Ο 6的每一過渡(舉例而言,過 渡3 a和3 b )之訊號8 0 2中的第三過渡之時序對此技 術的成功並非關鍵。亦即,由於閘極電感器會將閘極電壓 自然地推升至所需狀態且從每一閜極驅動裝置的汲極至源 極之寄生二極體會短暫地箝位閘極驅動電壓(亦即,訊號 802),所以,FET 108和110的閘極電壓會 自然地過渡至所需狀態。亦即,當電路的閘極共振中的能 量衰減時,閘極電壓會自然地衰減至所需位準。換言之, 訊號8 0 2的第三過渡(舉例而言,過渡3 a及3 b )之 時序不需爲關鍵參數。 也應注意,訊號8 0 2中的脈沖寬度無須受控制以致 於閘極訊號8 0 4會安定於所需邏輯位準而無振盪。亦即 ,可考慮允許少量的超量及共振振盪之實施例。雖然,這 些實施例並非如上述實施例般爲最佳的,但是,仍可從本 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -27- 502494 A7 B7 五、發明説明(25) 發明的技術獲利,舉例而言,共振振盪可限制於僅有一些 (其至是部份)週期。 本發明的另一特定實施例顯示於圖9 a中。所顯示的 簡化圖係表示技術的一般應用性。將參考圖9 b,說明圖 9 a的電路之操作。在本實施例中,使用浮動切換開關 9 0 2及電感器L 1以將切換開關9 0 4的閘極電容C g 切換至所需位準。驅動訊號由另一電容器C 1所代表,能 量會在二電容器之間交換而無超調。根據使用CM〇S技 術製造的電路之特定實施例,可使用如圖9 C所示之背對 背nM〇S及pMOS裝置,實施浮動切換開關9 0 2。 根據本實施例,簡單地藉由閉接,接著在適當時間斷 開切換開關9 0 2,在節點B可取得所需邏輯位準。如圖 9 b所示,當切換開關9 0 2在t 1閉接時,節點A處的 電壓,亦即3伏特,會開始衰減並經由電感器L 1以將節 點B處的電壓(亦即,橫跨閘極電容C g之電壓)充電。 然後,在適當時間t 2,再度斷開切換開關9 0 2 ,以確 保節點B達到所需位準。在切換開關9 0 2斷開時的適當 時間t 2係節點A及B達到交錯尖峰之時間。c g及C 1 之比例會影響節點A和B的訊號擺幅。亦如同上述,可採 用類似技術以適應地決定C 1與C g的比例及C 1的初始 充電。 此外,使用在適當時間將節點B連接至高或低(舉例 而言,3或0伏特)之切換開關,將節點B的電壓位準「 截掉頂部」。舉例而言,此技術可用以衰減與切換開關 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) (請先閱讀背面之注意事項再填寫本頁j -訂- 經濟部智慧財產局員工消費合作社印製 -28- 502494 A7 _ B7 五、發明説明(26) 9 0 2有關的串聯電阻之衰減效果。此外,藉由適當的偏 移C 1和C g的個別値,可克服切換開關9 0 2導入的電 阻性衰減。亦即,C 1可製成充份地大於C g以偏移共振 的振幅,藉以克服切換開關9 0 2造成的衰減。 雖然參考特定實施例,揭示發明,但是,應瞭解此處 所述的技術可應用至比此處所述的一些應用更寬廣的應用 。舉例而言,圖1係顯示本發明的帶通實施例,舉例而言 ,用於蜂巢式電話技術中的R F放大器。但是,本發明顯 然可應用至其它型式的放大器,舉例而言,基頻帶音頻放 大器。因此,本發明的範圍不應侷限於此處所述的實施例 ,而是應由申請專利範圍所決定。 現在將參考圖1及1 0,說明本發明的特定實施例。 亦即,圖1 0的電路與圖1的放大器一起使用以實現此處 所述的時計產生技術之實施。根據發明之特定實施例,從 節點A及B處的至少一共振,產生用於A / D轉換器 1 〇 4之時計(亦即,f s )。根據更特定的實施例, i s係部份地由節點A處的共振提供並部份地由節點B處 的共振提供。圖1 0係顯示可達成此目的之一實施。 比較器1 0 0 2會將節點A處的電壓(亦即,V a )與 小於功率輸出級的正電源之電壓(較佳地爲接地)比較。 當節點A以3 · 6 GHz在接地之上及之下共振時,產生 3 · 6 G Η z的時計。同樣地,比較器1 〇 〇 4會將節點 Β的電壓(亦即,V β )與接地以上的電壓(或負軌跡)相 比較,較佳地係與正電源比較,藉以當節點Β共振時,產 本紙張尺度適用中國國家標準(CNS ) Α4規格(210Χ297公釐) (請先閲讀背面之注意事項再填寫本頁) 訂 經濟部智慧財產局員工消費合作社印製 29· 502494 A7 B7 五、發明説明(27) (請先閲讀背面之注意事項再填寫本頁) 生3 · 6GHz之時計。乘法器1 008會在其輸入之間 選取,以回應控制邏輯(未顯示)所產生的控制訊號,根 據特定實施例,控制邏輯係取決於來自比較器1 0 4之決 定。以此方式,乘法器1008會產生時計訊號f s。當 電路首先斷開時,藉由產生使共振之一進行之脈沖,可啓 動時計訊號。根據一實施例,此可由選取環形振盪器 1006之乘法器1008完成。 此時計產生技術之一有利結果係由於時計至少部份地 由節點A及B處的共振所產生,所以,當這些共振上下移 動時(舉例而言,導因於反射及處理變異),輸至比較器 時計會以對應的方式上下移動。亦即,由A / D轉換器 1 0 4及閘極驅動電路1 0 6產生的閘極邊緣會比使用獨 立地產生之A / D時計更密切地符合輸出級共振的時序。 這會使切換邊緣較佳地對齊共振,以增進效率。 經濟部智慧財產局員工消費合作社印製 此外,導因於共振上下移動之方式之A/D轉換器時 計上的樣式相依顫動會將取樣比率有效地擾頻並將取樣頻 率相依音調擾亂成隨意雜訊,藉以消除輸出功率頻譜中取 樣頻率有關之不需要諧波。事實上,根據本發明,A / D 轉換器時計的「高頻振動」,與如何產生時計無關(舉例 而言,與自我計時無關),是依受控方式有意地導入的, 以擾亂取決於取樣頻率之雜訊音調。 再參考圖10及根據發明之更特定實施例,環式振盪 器1 0 0 6也可包含於時計產生電路中作爲增加的時計訊 號源。由於與輸出共振電路有關的衰減電阻可能高至足以 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -30- 502494 A 7 B7 五、發明説明(28) (請先閱讀背面之注意事項再填寫本頁) 造成共振振盪充份地衰減以致於它們不再觸動比較器 1 0 0 2和1 0 0 4且時計(因此閘極訊號)上鎖,所以 此點是需要的。因此,在從共振節點之一產生一些脈沖之 後(可能從與輸出級的共振電路相關之衰減電阻中依經驗 決定),多工器1008會受控以從環形振盪器選取輸入 作爲時計訊號直至另一節點上的共振開始。根據特定實施 例,環形振盪器會與共振振盪同步地啓動及停止,以致於 時計源之間的移交會平順。 經濟部智慧財產局員工消費合作社印製 根據特定實施例,使用此處所述的技術,增強上述參 考之對應申請案中所述之多位準切換發明,以提供自生時 計給多個獨立的時計A / D轉換器中的每一 A / D轉換器 。現在參考圖1 1,說明此實施例的特定實施例,圖1 1 係顯示R F帶通雜訊頻譜整型放大器1 1 〇 〇,其也是根 據上述中倂入參考之美國專利號5,777,5 12中所 揭示的技術設計。R F放大器1 1 〇 〇包含選頻網路 1 1 0 2,其使用連續時間回饋,將經過調變的r f輸入 施以雜訊頻譜整型。根據特定實施例,網路1 1 〇 2包括 至少一共振器,其係設計成使9 0 0 Μ Η z爲中心的頻帶 通過之轉換函數。 一 A/D轉換益1 1 〇 4 a及1 1 〇 4 b會以額定的 取樣頻率f S (亦即f S 1及f S 2 ),使用獨立產生的 時計訊號,將經過雜訊譜頻整型的R F訊號轉換成數位資 料,根據特定實施例,取樣頻率爲3 · 6 G Η z。根據一 實施例,A/D轉換器ll〇4a及ll〇4b包括用以 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -31 · 502494 A7 B7 五、發明説明(29) 實施三位準切換之二比較器。 閘極驅動電路1 1 〇 6 a及1 1 〇 6 b會分別從A/ (請先閱讀背面之注意事項再填寫本頁) D轉換器1 1 〇 4 a和1 1 〇 4 b取得脈沖列,並產生閘 極驅動以用於它們的電晶體對,亦即F E T ll〇8a 和 1110a 或 FET 1108b 及 1110b。每一 對電晶體具有分別導因於共振電路1 1 1 1和1 1 1 1 a 之二分別共振。亦即,包括FET 1108和1110 之功率級在節點A和B具有分別的共振,而包括F E T 1 1 0 8 a和1 1 1 〇 a之級在節點A ’和B,具有分別的共 振。 連續時間回饋會經由回饋路徑1 1 1 2和加法器 1 1 1 3,fen供結运頻網路1 1 〇 2。功率級的輸出訊號 會傳送給匹配網路1 1 1 4 ,匹配網路1 1 1 4會將輸出 R F訊號傳送給天線1 1 1 6以用於傳送。在另一實施例 中,加法器1 1 1 3係匹配網路1 1 1 4的部份。 經濟部智慧財產局員工消費合作社印製 具有用於A/D轉換器11〇4a和11〇4b之二 較器允許數位資料具有三量化狀態,亦即,三位準切換, 而非二位準。根據二量化狀態,可能會有很多造成高驅動 損失之訊號過渡。相對地,根據三狀態,當無訊號輸出以 避免此不需要的切換損失時,可選取“ 〇 ”狀態。 用於個別比較器1 1 0 4 a及1 1 〇 4 b之時計可由 獨立的光源產生。根據發明的特別實施例,時計訊號係根 據本發明的技術獨立地產生。亦即,根據發明之特定實施 例,會從節點A、B、A ’和B ’處的至少一些共振,產生用 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -32- 502494 A7 B7 五、發明説明(3〇) (請先閲讀背面之注意事項再填寫本頁) 於A/D轉換器1 1 04a及1 1 〇4b之時計(亦即, f s 1和f s 2 )。根據更特定的實施例,f s 1係部份 由節點A處的共振及部份由節點B處的共振所提供,而 f s 2係部份由節點A ’處的共振及部份由節點b ’處的共振 所提供。圖1 0係顯不可完成f 1及f 2的產生之一實施 。但是,將瞭解可以以其它方式產生這些時計訊號。 除了切換頻率相依雜訊之隨機化之外,尙有其它導因 於二時計的獨立本質之雜訊優點。此與平均時計頻率 f s 1及f s 2典型上相差某相當固定的量有關,此某相 當固定的量基本上是放大器遭遇之最小的重覆取樣率。亦 即,i s 1與f s 2之間的差造成「有效的」取樣頻率, 其遠高於f s 1或f s 2。結果,輸出雜訊頻譜中任何不 希望的音調或「輻射體」會因此非常高的「有效的」取樣 頻率而被移出有用頻帶。 經濟部智慧財產局員工消費合作社印製 根據特定實施例,f s 1與f s 2之間的差異是刻意 導入的並受控,以得到此效果的優點。根據時計f s 1及 i s 2從穩定的獨立源而非共振節點產生之更特定實施例 ,f s 1與f s 2之間的差異會受控而取得此優點。即使 在f s 1與f s 2從單一源導出之情形,差異仍會被導入 以取得優點並仍維持在發明之範圍內。 雖然參考發明之特定實施例以特別地說明發明,但是 習於此技藝者應瞭解在不悖離發明的精神或範圍之下,所 揭示之實施例之形式及細節可以有變化。舉例而言,僅由 節點A和B中的一者結合環形振盪器以產生時計訊號之實 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公羡) -33- 502494 A7 B7 五、發明説明(31) (請先閱讀背面之注意事項再填寫本頁) 施例。由於施加至功率級的輸入之閘極脈沖可能較不匹配 其它共振的時序,所以,此實施例可能較不需要。但是, 此實施例顯然在本發明的範圍內。 也應注意,雖然此處參考帶通(舉例而言,RF)實 施,以說明發明,但是,本發明也可應用至其它放大器配 置,舉例而言,基頻帶音頻放大器。因此,發明之範圍應 由申請專利範圍決定。 經濟部智慧財產局員工消費合作社印製 圖1 2係根據發明的特定實施例之調變器迴路 1 2 0 0之方塊圖。調變器級1 2 0 2可爲廣泛的調變器 型式中的任意者,這些調變器包含脈寬調變器、或是諸如 美國專利號5,7 7 7,5 1 2中所揭示之過取樣混合訊 號調變器。根據特定實施例,調變器級1 2 0 2包括多個 並聯及/或串聯之濾波器。切換級1 2 0 4從調變器 1 2 0 2的輸出接收控制訊號。切換級1 2 0 4可爲多種 切換開關配置及功率等級中的任意者。切換級1 2 0 4也 可爲反相或非反相。對於反相切換級而言,反相器會插入 於回饋濾波器1 2 0 8之前的調變器輸出回饋路徑中。根 據多種適用於特別應用中的所需輸出訊號內容之技術中的 任意者,也可實施輸出濾波器1 206。 如同以往設計般,負回饋可從切換級1 2 0 4的輸出 及/或輸出濾波器1 2 0 6 ,經由總合接頭提供給調變器 1 2 0 2。爲了說明此點,假定接頭1 2 1 0爲理想的, 因此,未顯示增加的回饋電路。但是,將瞭解必須使用諸 如電阻器等真正的電路元件以實施總合接頭1 2 1 0。因 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) •34- 502494 A7 B7 五、發明説明(32) (請先閱讀背面之注意事項再填寫本頁) 此,將於下參考圖4 一 6,說明顯示總合接頭1 2 1 0的 某些真正實施例之發明的特定實施例。來自切換級及輸出 濾波器之回饋的頻率內容均在有用之迴路的頻率範圍之內 及之外。 與舊有的設計不同,調變器迴路1 2 0 0也採用經過 回饋濾波器1 2 0 8來自調變器1 2 0 2的輸出之回饋, 以進一步增進迴路穩定度。但是,爲了避免發明背景中所 討論的輸出逼真度之不利結果,此回饋會由濾波器 1 2 0 8濾波,以致於其頻率內容大部份在有用的調變器 迴路頻率範圍之外。因此,穩定度會增加,但不會損失逼 真度。 經濟部智慧財產局員工消費合作社印製 圖1 3係根據發明的更特定實施例設計之基頻帶調變 器迴路1 3 0 0之方塊圖。調變器1 3 0 2、切換級 1 3 0 4、及輸出濾波器1 3 0 6實際上以類似於上述參 考圖1 2所述之對應的迴路元件之方式操作。回饋會經由 包括電阻器1 3 1 2和1 3 1 4之除法器網路,從切換級 1 3 0 4的輸出提供給調變器1 3 0 2的回饋輸入。回饋 也會經由濾波器電容器1 3 0 8和串聯電阻器1 3 1 0, 從其邏輯輸出,提供給調變器1 3 0 2的回饋輸入。選取 電容器1 3 0 8的値,以將有用之基頻帶範圍中的邏輯輸 出訊號內容衰減,並使較高的頻率通至回饋路徑以與經過 衰減的切換級訊號結合。藉由此配置,可顯著地增加延遲 容許度。 但是,由於經過濾波的調變器邏輯訊號未與使用理想 本紙張尺度適用中國國家標準(CNS ) A4規格(210Χ297公董) ~" 502494 A7 B7__ 五、發明説明(33) (請先閱讀背面之注意事項再填寫本頁) 濾波器及理想總合器之回饋路徑結合,所以,會由真實電 路元件(亦即,電容器1308和電阻器1310)導入 額外的極及零至迴路的轉換函數中。將調變器1 3 0 2的 邏輯輸出視爲接地並觀察電容器1 3 0 8藉由低通濾波切 換級輸出以將延遲加至迴路,則可從網路分析透視圖中看 到此點。此延遲代表藉由使用經過濾波的調變器輸出作爲 回饋而可能造成的性能增進之限制。因此,將於下說明發 明的其它實施例,其增進圖1 3的迴路,以說明此限制。 經濟部智慧財產局員工消費合作社印製 圖1 4係根據發明之又更特定實施例設計之基頻帶調 變器迴路1400的方塊圖。調變器1402、切換級 1 4 0 4、及輸出濾波器1 4 0 6實際上以類似於上述參 考圖1 2說明之對應的迴路元件之方式操作。根據迴路 1 3 0 0的對應元件,回饋會經由包括電阻器1 4 1 2和 1 4 1 4的除法器網路,從切換級1 4 0 4的輸出提供給 調變器1 4 0 2的回饋輸入。回饋也會經由濾波器電容器 1 4 0 8及串聯電阻器1 4 1 0,從其邏輯輸出提供給調 變器1 40 2的回饋輸入。選取電容器1 40 8的値以衰 減有用的基頻帶範圍中的邏輯輸出訊號內容並使較高頻率 通至回饋路徑以與經過衰減的切換級訊號結合。 從切換級1 4 0 4的輸出至節點1設置增加的電容器 1416,其會抵消電容器1408導入切換級的回饋路 徑之極。再度藉由視調變器1 4 0 2的邏輯輸出爲接地及 分析來自切換級1 4 0 4的輸出之回饋路徑,可從網路分 析透視圖瞭解此點。假使電阻器1 4 1 2與1 4 1 4的比 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -36- 502494 A7 B7 五、發明説明(34) 例等於電容器1 4 1 6與1 4 0 8的比例,則在節點]_處 看到的切換級輸出之衰減與節點2處看到的相同,亦即, 電壓位準應相同。由於節點1及2處的電壓相等,所以, 無電流流經電阻器1 4 1 0,因此,電阻器1 4 1 0可被 視爲開路,可從切換級回饋路徑有效地移除電容器 1 4 0 8及其對應極。這將使得調變器迴路相當穩定以及 容許延遲。 圖1 5係根據發明的又更特定實施例設計之調變器迴 路1 500的方塊圖。調變器1 502、切換級1504 、及輸出濾波器1 5 0 6實際上以類似於上述參考圖1 2 所述的對應迴路元件之方式操作。電阻器1 5 1 0、 1512及1514以及電容器1508及1516實際 上以類似於上述參考圖1 4所述之對應的迴路元件之方式 操作。事實上,圖1 5的實施例包含圖1 4的所有電路加 上從輸出濾波器1 5 0 6至調變器級1 5 0 2的回饋輸入 之回饋路徑,增加的回饋路徑係以電阻器1 5 2 0實施。 電容器1 5 2 2也包含從濾波器1 5 0 6的輸出至節點1 之電容器1522,用抵消電容器1508和1516導 入的新回饋路徑中的極。如同上述參考圖1 4之實施例所 述般,假使電阻器1 5 2 0與並聯電阻器1 5 1 2之電阻 器1 5 1 4的比例同於並聯電容器1 5 1 6之電容器 1 5 0 8與電容器1 5 2 2之比例,則無電流流入電阻器 1 5 1 0,因此,可從任一回饋路徑有效地移除電容器 1 5 0 8° 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) (請先閲讀背面之注意事項再填寫本頁) 訂 經濟部智慧財產局員工消費合作社印製 -37- 502494 A7 B7 五、發明説明(奶) (請先閱讀背面之注意事項再填寫本頁) 圖1 6係根據發明設計之調變器迴路1 6 0 0的一般 實施例之方塊圖。其以4埠網路1 6 0 8取代先前所述的 實施例之回饋濾波器及訊號混合電路,調變器級1 6 0 2 、切換級1 6 0 4、及輸出濾波器級1 6 0 6會連接至4 埠網路1 6 0 8。網路1 6 0 8會取得調變器邏輯輸出訊 號並衰減其在調變器迴路1 6 0 0的有用範圍內之頻率內 容。其也會結合所造成的訊號與切換及濾波器級之回饋輸 出,以作爲調變器級1 6 0 2的回饋。可視用以實施調變 器迴路1 6 0 0之技術而以不同方式實施網路1 6 0 8。 此外,對於具有振幅相依穩定度之調變器而言(舉例 而言,過取樣及脈寬調變器),可能需要控制與來自功率 及/或濾波器級回饋有關的調變器回饋之量値,以維持最 佳穩定度點。因此,在回饋濾波器(包含4埠網路 1 6 0 8 )之前提供選加的增益控制1 6 1 0,以允許調 變器邏輯輸出訊號與功率輸出之增益匹配。根據特定實施 例,使用具有可變輸出振幅之數位對類比轉換器(D A C )增益控制1 6 1 0。 經濟部智慧財產局員工消費合作社印製 雖然已參考發明之特定實施例,顯示及說明發明,但 是,習於此技藝者應瞭解,在不悖離發明的精神及範圍之 下,所揭示的實施例之形式及細節可以改變。舉例而言, 參考圖1 3及1 4之上述特定實施例顯示使用類比元件( 舉例而言,電阻器及電容器)於濾波器及調變器輸出並結 合所造成的經過濾波之回饋訊號與來自切換級的輸出之回 饋路徑。但是,應瞭解,舉例而言,在調變器級係使用數 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -38- 502494 A7 B7 五、發明説明(36) (請先閲讀背面之注意事項再填寫本頁) 位技術實施之情形中,用以將回饋訊號濾波及結合之電路 及技術可爲數位式的。同樣地,藉由混合訊號調變器、混 合訊號電路和技術可用以實施回饋,舉例而言,混合運算 放大器的實際接地端之訊號。 也應瞭解,範圍中調變器輸出的元件之衰減的變化程 度是否適當,係取決於調變器的內容及有用的調變器迴路 頻率範圍中功率輸出之程度,以及迴路的輸出可忍受之惡 化程度。亦即,無須完全地拒絕有用的迴路頻率範圍內之 調變器級輸出中的頻率成份,以維持合理的逼真度及維持 在發明的範圍之內。 經濟部智慧財產局員工消費合作社印製 圖1 7係顯示根據本發明以及美國專利號 5,7 7 7,5 1 2中所揭示的技術所設計之R F帶通雜 訊譜頻整型放大器1 7 0 0。舉例而言,此放大器可用於 諸如無線電話、呼叫器、網路裝置、等等中。r F放大器 1700包含選頻網路1702,選頻網路1702使用 連續時間回饋以將經過調變的R F輸入施以雜訊譜頻整型 。根據特定實施例,網路1 7 0 2包括至少一共振器級, 此至少一共振器級具有設計成以9 0 0 Μ Η z爲中心的頻 帶可通過之轉換函數。當然,應瞭解根據發明設計之雜訊 譜頻整型放大器的帶通頻率範圍會以諸如1 . 8 G Η ζ、 2 · 4GHz、等多個頻率中的任一'頻率爲中心,且本發 明不限於任何特定範圍或範圍組。 A / D轉換器1 7 0 4會使用取樣頻率f s,將經過 雜訊頻譜整型的R F訊號轉換成數位資料,根據特定實施 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) -39 - 經濟部智慧財產局員工消費合作社印製 502494 A7 B7 五、發明説明(37) 例,取樣頻率f s爲3 · 6 G Η z。根據一實施例,A / D轉換器1 7 0 4包括比較器。 閘極驅動電路1 7 0 6會從A/D轉換器1 7 〇 4取 得脈沖列並產生用於放大器1 7 0 0的功率輸出級之每— F E T 1 7 0 8和1 7 1 0之閘極驅動。所示之輸出功 率級包含電怠器LI、L2及L3、以及電容器C1 °當 FET 1708及1710中對應的一者閉接時,此配 置會在節點A及B處分別產生二分別的共振。經由回饋路 徑1 7 1 2 ,提供連續時間回饋給選頻網路1 7 0 2。功 率級的輸出訊號會傳送給匹配網路1 7 1 4,匹配網路 1 7 1 4會將輸出R F訊號傳送給天線1 7 1 6以用於傳 送。如同將瞭解的及根據本發明,回饋路徑1 7 1 2可源 起於匹配網路1 7 1 4之內。 放大器1 7 0 0的迴路型典上最佳地係以等於1 / f s (亦即,1/3 · 6GHz)之迴路延遲操作。爲了 補償與輸出功率級有關的任何增加及不需要的延遲,放大 器1 7 0 0包含通至選頻網路1 7 0 2之另一回饋路徑 1718,選頻網路1702包含與網路1702具有相 同中心點之陷波濾波器1 7 2 0。回饋的源極顯示成如同 A / D轉換器1 7 0 4的輸出。但是,將瞭解此節點與通 至回饋源起之功率級的輸入之間會有數個其它點。因此, 本發明的範圍不應受限於所示之實施例。而是,如同看待 此處所述的所有實施例一般,本發明的範圍應以下述申請 專利範圍解釋。 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) (請先閱讀背面之注意事項再填寫本頁) 、11 -40- 502494 A7 B7 五、發明説明(38) 雖然以蜂巢式電話應用說明本發明,但是,顯然發明 可以以多種內容及其它實施例,實施發明。舉例而言,本 發明的R F放大器可用於需要有效率的r f放大器之蜂巢 式電話基地台或任何其它應用。本發明的數位放大器也可 用於諸如1, 8GHz或3.0GHz或更高的頻率之 R F傳送。在這些實施例中,藉由依據上述比例增加取樣 頻率(f s ),可取得更高的傳送速率。也將瞭解,本發 明的放大器配置及輸出切換級之用途不限於R F應用。 電晶體T1及T2也可爲一些不同型式的裝置,包含 MESFET、 HBT、 CMOS、或 NMOS,並可由 一些不同的製程及材料(包含GaAs、SiGe、或標 準矽)製成。因此,可瞭解此處所提供的實施例僅爲舉例 說明,發明之真正範圍及精神應由申請專利範圍決定。 (請先閲讀背面之注意事項再填寫本頁) 訂 經濟部智慧財產局員工消費合作社印製 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) 41 -Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. 5. Description of the invention (21) The rate f S (ie, fsi and fs 2), using an independently generated timepiece signal, will pass through the noise signal that is shaped by the noise spectrum. In a specific embodiment, the nominal sampling frequency f sm3 _ 6eHZ. According to the embodiment, the A / D converters 604a and 6o4b include two comparators for implementing three-level switching. The squeegee driving circuit 6 0 6 a and 6 0 6 b will obtain the pulse train from the A / D converters 60 4 a and 6 0 4 b, respectively, and generate a pair of transistors for the transistor (ie, FET 608a and 6). 〇3 or 1?] 2 丁 6 0 8 b and 6 1 〇 b) gate drive. Each pair of transistors has two resonances due to the resonator circuits 6 1 1 and 6 1 a respectively. That is, the power stages including F E T 6 0 8 and 6 1 Q have respective resonances at nodes a and B, and the stages including FETs 608a and 610a have respective resonances at nodes A 'and B'. The continuous time feedback is provided to the frequency selection network 6 0 2 via the feedback path 6 1 2 and the adder. The output signal of the power stage will be sent to the matching network 6 1 4 and the matching network 6 1 4 will send the output RF signal to the antenna 6 1 6 for transmission. Having two comparators for converters 604a and 604b will allow digital signals to have three quantization states, that is, three-level switching, rather than two-level switching. For example, depending on the two-level quantization state of the amplifier 4 3 4, there may be many signal transitions that cause high drive losses. In contrast, according to the three states, when there is no signal output to avoid this unnecessary switching loss, the state of "0" can be selected. Figure 7 includes Figures 7A, 7B, and 7C, which are detailed diagrams showing switching levels, according to a specific implementation of the present invention. National Standard (CNS) A4 Specification (210x297 mm) ---------% II (Please read the notes on the back before filling this page) Order • Γ. -24- 502494 A7 B7 Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. 5. Description of the invention (22 〇 1 1 example, which can be used by the RF amplifier of Figure 6). However, it should be understood that it does not depart from the scope of invention The switching stage of Figure 7 below and the switching stage of the amplifier 1 I of Figures 5 and 6 can be used in different amplifier types. That is, the ^ -V of the present invention. 1 1 The configuration of the switching stage is applicable outside the RF range. Many applications. Read first. 1 1 Although the present invention is described by using a cellular phone application, it should be understood on the back 1 that the invention can be implemented with a wide range of connotations and other embodiments. Note that the RF of the invention The amplifier can be used in a cellular telephone base station or any other application that requires an efficient RF amplifier. 1 The digital amplifier of the present invention • The I amplifier can also be used such as 1 • 8 G Η Z or 3 • 0 G Η Z or higher frequency page 1 1 RF transmission. In this itb embodiment, a higher transmission rate can be achieved by increasing the sampling rate (f S) according to the above ratio 1 1. Transistors T 1 and 1 IT 2 can also be a & different type of device, including Μ ESF Ε 丁 定 I Η BT CM 〇S, or N Μ 〇S and can be made by some different processes 1 1 I and materials (including G a AS, S i Ge or standard silicon) is made 〇 Because 1 1 Therefore, you can understand that the embodiment provided here is only an example, the true scope of the invention 1 | The positive scope and spirit should be determined by the scope of the patent claim 0 # 1 According to a specific embodiment The gate driving circuit 10 6 of the noise spectrum shaping amplifier of FIG. 1 operates in accordance with the principles of the present invention. Referring to FIG. 8 a shows the rising and falling edges of the gate signal for one of the 1 1 FETs. At this point, the gate signal 8 0 2 is a gate driving circuit 1 0 6 and the signal 1 I Drfe · Wl 8 0 4 is located at the gate of the FET. 0 The signal 8 0 4 shown is used for either or both of the 1 1 IFET 10 8 and 1 1 0. One of them is basically 1 1 The complement of the other 0 is used for the gate drive circuit 1 0 6 One possible implementation system 1 1 is shown in Figure 8b 〇 But it should be understood that it does not depart from the scope of the invention 1 1 1 This paper scale applies to China National Standard (CNS) A4 Specification (210X297mm) -25- 502494 A7 B7 V. Description of Invention (23) In this embodiment and other embodiments described herein, a variety of driving circuits can be used. (Please read the notes on the back before filling this page) According to an embodiment, just enough energy is introduced into the resonance at the gate of the FET, so that the signal at the gate 8 0 4 will be at the required input logic bit quasi. The width of the initial pulse used for each logic transition in the gate signal 802 determines this energy and is controlled by the gate drive circuit 106. That is, the gate driving circuit 106 is used to receive the pulse train 8 0 6 from the A / D converter 104 (for example, it can be a comparator) and generate a gate signal. Each positive and negative transition of the pulse train of the D converter 104 corresponds to the corresponding transition sequence shown in Fig. 8a. Therefore, for the transition from the logic “0” to the logic “1” of the signal 8 0 6, the gate driving circuit 10 6 will generate in the signal 8 0 2 and cause the signal 8 0 4 to be at the beginning of the width W 1 of the logic 1 pulse. Then, before the next transition of the signal 8 0 · at a later point, the signal 8 2 is brought to a high level again (ie, the third transition). At this point, a similar sequence of transitions takes place to bring the transition to logic "0". As you can see, printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs, this will not only prevent resonance from oscillating, but also solve the problem of overshoot / downshoot. Since some resistors are connected in series with the input of each FET, the width W of each initial transition pulse in the signal 802 needs to be wide enough to overcome the attenuation factor introduced by this resistor. Otherwise, the level of signal 804 will never reach the required logic level. On the other hand, it is also desirable that the width W is not too wide to avoid significant overshoot or oscillation. Therefore, according to a specific embodiment of the present invention, adaptive feedback control can be provided, which will supervise the signal 8 0 4 ^ Paper size applies Chinese National Standard (CNS) A4 specification (210X29? Mm) ~ -26-502494 A7 B7 5 Explanation of the invention (24) To the extent that the required logic level is reached or overcome, and use this information to adjust the timing of the transition pulse of the signal 8 0 2 through the gate drive circuit 1 06. (Please read the note on the back first Please fill in this page again) According to another specific embodiment, the adaptive loop includes at least one comparator for measuring whether the signal level at the gate of the FET is too high or too low, and using a DAC to adjust the gate driving circuit 1 〇 The bias current of the one or more delay lines adjusts the timing of the transition pulse of the signal 802 to correct the excess or deficiency. According to different embodiments, states 2a and 2b can be achieved by closing the driving circuit. Inductor current will be provided for transitions 2a, 2b, 3a and printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 3b. This means that the timing of the third transition in signal 8 0 2 for each transition of signal 8 0 6 (for example, transitions 3 a and 3 b) is not critical to the success of the technology. That is, since the gate inductor naturally pushes the gate voltage to the desired state and the parasitic diode from the drain to the source of each pole driving device will temporarily clamp the gate driving voltage (also That is, the signal 802), so the gate voltages of the FETs 108 and 110 naturally transition to a desired state. That is, when the energy in the gate resonance of the circuit decays, the gate voltage naturally decays to the required level. In other words, the timing of the third transition of the signal 8 02 (for example, transitions 3 a and 3 b) need not be a key parameter. It should also be noted that the pulse width in the signal 802 need not be controlled so that the gate signal 804 will settle to the required logic level without oscillation. That is, an embodiment that allows a small amount of excess and resonance oscillations can be considered. Although these embodiments are not as optimal as the above-mentioned embodiments, the Chinese paper standard (CNS) A4 specification (210X297 mm) can still be applied from the paper size. -27- 502494 A7 B7 V. Description of the invention (25 ) The technology of the invention is profitable. For example, resonance oscillations can be limited to only a few (to some extent) periods. Another specific embodiment of the invention is shown in Fig. 9a. The simplified diagram shown shows the general applicability of the technology. The operation of the circuit of Fig. 9a will be explained with reference to Fig. 9b. In this embodiment, a floating changeover switch 902 and an inductor L 1 are used to switch the gate capacitance C g of the changeover switch 904 to a desired level. The driving signal is represented by another capacitor C1, and the energy will be exchanged between the two capacitors without overshoot. According to a specific embodiment of a circuit manufactured using CMOS technology, a back-to-back nMOS and pMOS device as shown in FIG. 9C may be used to implement a floating changeover switch 902. According to this embodiment, simply by closing, and then opening the switch 902 at an appropriate time, the required logic level can be obtained at the node B. As shown in Figure 9b, when the switch 9 0 2 is closed at t 1, the voltage at node A, that is, 3 volts, will begin to decay and pass the inductor L 1 to reduce the voltage at node B (that is, , The voltage across the gate capacitance C g). Then, at an appropriate time t 2, the switch 9 0 2 is turned off again to ensure that the node B reaches the required level. The appropriate time t 2 when the switch 9 0 2 is turned off is the time when the nodes A and B reach the staggered peaks. The ratio of c g and C 1 will affect the signal swing of nodes A and B. Also as described above, similar techniques can be used to adaptively determine the ratio of C 1 to C g and the initial charge of C 1. In addition, use a toggle switch that connects Node B to high or low (for example, 3 or 0 volts) at the appropriate time to "truncate the top" of the voltage level of Node B. For example, this technology can be used to attenuate and switch. This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) (Please read the precautions on the back before filling out this page. J-Order-Intellectual Property Bureau, Ministry of Economic Affairs Printed by the employee consumer cooperative -28- 502494 A7 _ B7 V. Attenuation effect of series resistance related to invention description (26) 9 0 2. In addition, by appropriately shifting individual 値 of C 1 and C g, the switching can be overcome The resistive attenuation introduced by the switch 9 0 2. That is, C 1 can be made sufficiently larger than C g to shift the amplitude of the resonance, thereby overcoming the attenuation caused by the switching switch 9 0 2. Although the invention is disclosed with reference to a specific embodiment However, it should be understood that the techniques described herein can be applied to a wider range of applications than some of the applications described here. For example, Figure 1 shows a bandpass embodiment of the present invention, for example, for the hive Amplifiers in conventional telephone technology. However, the present invention is obviously applicable to other types of amplifiers, for example, baseband audio amplifiers. Therefore, the scope of the present invention should not be limited to the embodiments described herein. Instead, it should be determined by the scope of the patent application. A specific embodiment of the present invention will now be described with reference to Figs. 1 and 10. That is, the circuit of Fig. 10 is used with the amplifier of Fig. 1 to implement the timepiece described herein Implementation of generation technology. According to a specific embodiment of the invention, a timepiece (ie, fs) for A / D converter 104 is generated from at least one resonance at nodes A and B. According to a more specific embodiment, is is provided in part by the resonance at node A and part is provided by the resonance at node B. Figure 10 shows one implementation that can accomplish this. Comparator 1 0 2 will apply the voltage at node A (Ie, V a) compared to the voltage of the positive power supply (preferably ground) which is less than the power output stage. When node A resonates above and below ground at 3 · 6 GHz, it generates 3 · 6 G 产生z timepiece. Similarly, the comparator 1004 compares the voltage at the node B (that is, V β) with the voltage above the ground (or negative trajectory), preferably with a positive power source, so that when When Node B resonates, the paper size of the production paper applies to the Chinese National Standard (CNS) Α4. (210 × 297 mm) (Please read the notes on the back before filling this page) Order printed by the Intellectual Property Bureau Employee Consumer Cooperatives of the Ministry of Economic Affairs 29 · 502494 A7 B7 V. Invention Description (27) (Please read the notes on the back before (Fill in this page) is a 3 · 6GHz timepiece. The multiplier 1 008 selects between its inputs in response to the control signal generated by the control logic (not shown). According to a specific embodiment, the control logic is determined by the comparator The decision of 104. In this way, the multiplier 1008 generates a time signal fs. When the circuit is first disconnected, the time signal can be started by generating a pulse that causes one of the resonances to proceed. According to an embodiment, this can be done by a multiplier 1008 that selects the ring oscillator 1006. One of the favorable results of the timepiece generation technology is that the timepiece is generated at least in part by the resonances at nodes A and B. Therefore, when these resonances move up and down (for example, due to reflection and processing variation), the The comparator timepiece moves up and down in a corresponding manner. That is, the gate edges generated by the A / D converter 104 and the gate driving circuit 106 will more closely match the timing of the output stage resonances than when the A / D timepiece generated independently is used. This will allow the switching edges to align better with resonance for increased efficiency. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. In addition, the pattern-dependent chattering on the A / D converter timepiece caused by the resonance moving up and down will effectively scramble the sampling ratio and disturb the sampling frequency dependent tone into random noise. To eliminate unwanted harmonics related to the sampling frequency in the output power spectrum. In fact, according to the present invention, the "high frequency vibration" of A / D converter timepieces has nothing to do with how the timepieces are generated (for example, has nothing to do with self-timing) and is intentionally introduced in a controlled manner to disrupt Noise tones at the sampling frequency. Referring again to Fig. 10 and according to a more specific embodiment of the invention, a ring oscillator 1 0 6 may also be included in the timepiece generating circuit as an added timepiece signal source. Because the attenuation resistance related to the output resonance circuit may be high enough for this paper size to apply the Chinese National Standard (CNS) A4 specification (210X297 mm) -30- 502494 A 7 B7 V. Description of the invention (28) (Please read the first Note that you need to fill in this page again) This causes the resonance oscillations to decay sufficiently that they no longer touch the comparators 1 0 2 and 1 0 4 and the timepiece (hence the gate signal) is locked, so this point is needed. Therefore, after generating some pulses from one of the resonance nodes (possibly determined from the attenuation resistance associated with the resonance circuit of the output stage), the multiplexer 1008 is controlled to select the input from the ring oscillator as the timepiece signal until the other Resonance at one node begins. According to a specific embodiment, the ring oscillator is started and stopped in synchronization with the resonance oscillation, so that the handover between the timepiece sources is smooth. Printed by the Consumer Property Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. According to certain embodiments, the technology described here is used to enhance the multi-level switching invention described in the corresponding application of the above reference to provide self-generated timepieces to multiple independent timepieces. Each A / D converter in the A / D converter. A specific embodiment of this embodiment will now be described with reference to FIG. 11. FIG. 11 shows an RF bandpass noise spectrum shaping amplifier 1 1OO, which is also based on US Patent No. 5,777, which is incorporated herein by reference. 5 The technical design disclosed in 12. The R F amplifier 1 1 0 0 includes a frequency selective network 1 102, which uses continuous time feedback to apply a modulated r f input to a noise spectrum integer. According to a specific embodiment, the network 1 102 includes at least one resonator, which is a transfer function designed to pass a frequency band centered at 900 MHz. A A / D conversion benefit 1 1 〇 4 a and 1 1 〇 4 b will use the rated sampling frequency f S (ie f S 1 and f S 2), using the independently generated timepiece signal, will pass the noise spectrum frequency The integer RF signal is converted into digital data. According to a specific embodiment, the sampling frequency is 3 · 6 G Η z. According to an embodiment, the A / D converters 1104a and 1104b include the Chinese paper standard (CNS) A4 specification (210X297 mm) for this paper size -31 · 502494 A7 B7 V. Description of the invention (29) Implement a two-level quasi-switching two comparator. The gate drive circuits 1 1 〇6 a and 1 1 〇6 b will obtain the pulse train from A / (please read the precautions on the back before filling in this page) D converter 1 1 〇4 a and 1 1 〇4 b And generate gate drivers for their transistor pairs, namely FETs 1108a and 1110a or FETs 1108b and 1110b. Each pair of transistors has two resonances respectively due to the resonance circuits 1 1 1 1 and 1 1 1 1 a. That is, the power levels including FETs 1108 and 1110 have respective resonances at nodes A and B, and the levels including F E T 1 1 0 8 a and 1 1 1 0 a have respective resonances at nodes A 'and B. The continuous time feedback will pass through the feedback path 1 1 1 2 and the adder 1 1 1 3, and the fen is used to end the carrier frequency network 1 1 02. The output signal of the power stage will be sent to the matching network 1 1 1 4, and the matching network 1 1 1 4 will send the output RF signal to the antenna 1 1 1 6 for transmission. In another embodiment, the adder 1 1 1 3 is part of the matching network 1 1 1 4. The Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs has printed two comparators for A / D converters 1104a and 1104b to allow digital data to have a three-quantity state, that is, three-level switching, not two-level . According to the binary state, there may be many signal transitions that cause high drive losses. In contrast, according to the three states, when there is no signal output to avoid this unnecessary switching loss, the state of "0" can be selected. The timepieces for the individual comparators 1 10 4 a and 1 104 b can be generated by independent light sources. According to a particular embodiment of the invention, the timepiece signal is generated independently according to the technology of the invention. That is, according to a specific embodiment of the invention, at least some resonances at the nodes A, B, A ', and B' will be generated to use the Chinese paper standard (CNS) A4 specification (210X297 mm) for this paper size -32- 502494 A7 B7 V. Description of the invention (30) (Please read the notes on the back before filling this page) At the time of A / D converter 1 1 04a and 1 1 〇4b (ie, fs 1 and fs 2) . According to a more specific embodiment, fs 1 is partly provided by resonance at node A and part is provided by resonance at node B, while fs 2 is partly provided by resonance at node A 'and partly by node b' The resonance provided here. Fig. 10 shows the implementation of one of the unfinished f 1 and f 2 generations. However, it will be appreciated that these timepiece signals can be generated in other ways. In addition to the randomization of switching frequency-dependent noise, there are other noise advantages due to the independent nature of the two-hour clock. This is related to the fact that the average timepiece frequencies f s 1 and f s 2 differ by some fairly constant amount, which is basically the smallest repeated sampling rate encountered by the amplifier. That is, the difference between i s 1 and f s 2 results in an "effective" sampling frequency that is much higher than f s 1 or f s 2. As a result, any undesired tones or "radiators" in the output noise spectrum will be shifted out of the useful band due to the very high "effective" sampling frequency. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs According to a specific embodiment, the difference between f s 1 and f s 2 is deliberately introduced and controlled to obtain the advantages of this effect. According to a more specific embodiment in which the timepieces f s 1 and is s 2 are generated from stable independent sources rather than resonance nodes, the difference between f s 1 and f s 2 can be controlled to achieve this advantage. Even in the case where f s 1 and f s 2 are derived from a single source, the differences will still be introduced to gain advantages and remain within the scope of the invention. Although the invention is specifically described with reference to specific embodiments of the invention, those skilled in the art should understand that the form and details of the disclosed embodiments may vary without departing from the spirit or scope of the invention. For example, only one of nodes A and B combined with the ring oscillator to generate the actual timepiece signal. The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X297 public envy) -33- 502494 A7 B7 V. Invention Note (31) (Please read the notes on the back before filling out this page). Since the gate pulses applied to the input of the power stage may not match the timing of other resonances, this embodiment may be less needed. However, this embodiment is clearly within the scope of the present invention. It should also be noted that although reference is made here to bandpass (for example, RF) implementations to illustrate the invention, the invention can also be applied to other amplifier configurations, such as baseband audio amplifiers. Therefore, the scope of the invention should be determined by the scope of the patent application. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs Figure 12 is a block diagram of a modulator circuit 1 2 0 0 according to a specific embodiment of the invention. The modulator stage 1 2 0 2 can be any of a wide range of modulators, including pulse width modulators, or such as disclosed in US Patent No. 5,7 7 7,5 1 2 Oversampling mixed signal modulator. According to a particular embodiment, the modulator stage 1 2 0 2 comprises a plurality of filters connected in parallel and / or in series. The switching stage 1 2 0 4 receives a control signal from the output of the modulator 1 2 0 2. The switching level 1 2 0 4 can be any of a variety of switching switch configurations and power levels. The switching stage 1 2 0 4 can also be inverted or non-inverted. For the inverting switching stage, the inverter is inserted into the modulator output feedback path before the feedback filter 1 2 0 8. The output filter 1 206 may also be implemented according to any of a variety of techniques suitable for the content of the desired output signal in a particular application. As in previous designs, negative feedback can be provided from the output of the switching stage 1 2 0 4 and / or the output filter 1 2 0 6 to the modulator 1 2 0 2 via the summation joint. To illustrate this, it is assumed that the connector 1 2 1 0 is ideal, and therefore, an added feedback circuit is not shown. However, it will be understood that a true circuit element such as a resistor must be used to implement the header 1 2 1 0. Because this paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) • 34- 502494 A7 B7 V. Description of the invention (32) (Please read the precautions on the back before filling this page) This will be referenced below Figures 4 to 6 illustrate a particular embodiment of the invention showing some real embodiments of the joint joint 1210. The frequency content of the feedback from the switching stage and output filter is within and outside the frequency range of the useful loop. Unlike the old design, the modulator circuit 1 2 0 0 also uses the feedback from the output of the modulator 1 2 0 2 through the feedback filter 1 2 0 8 to further improve the loop stability. However, in order to avoid the unfavorable results of the output fidelity discussed in the background of the invention, this feedback is filtered by the filter 128, so that most of its frequency content is outside the useful modulator loop frequency range. As a result, stability is increased without loss of fidelity. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs Figure 13 is a block diagram of a baseband modulator circuit 1 300 according to a more specific embodiment of the invention. The modulator 1 3 0 2, the switching stage 1 3 0 4, and the output filter 1 3 0 6 actually operate in a manner similar to the corresponding loop element described above with reference to FIG. 12. The feedback is provided from the output of the switching stage 1 3 0 4 to the feedback input of the modulator 1 3 0 2 via a divider network including resistors 1 3 1 2 and 1 3 1 4. The feedback is also provided to the feedback input of the modulator 1 3 0 2 from its logic output via the filter capacitor 1 3 0 8 and the series resistor 1 3 1 0. The capacitor 1 3 0 8 is selected to attenuate the content of the logical output signal in the useful baseband range, and pass the higher frequency to the feedback path to combine with the attenuated switching level signal. With this configuration, the delay tolerance can be significantly increased. However, because the filtered logic signal of the modulator is not ideal, the Chinese paper standard (CNS) A4 (210 × 297) is used in this paper size. &Quot; 502494 A7 B7__ V. Description of the invention (33) (Please read the back first (Please note this page and fill in this page again) The combination of the filter and the feedback path of the ideal combiner, so the real circuit components (ie, capacitor 1308 and resistor 1310) will be introduced into the extra pole and zero-to-loop transfer function . Considering the logic output of the modulator 1 3 0 2 as ground and observing the capacitor 1 3 0 8 switches the output of the stage by low-pass filtering to add delay to the loop. This can be seen from the network analysis perspective. This delay represents a limit on the performance gain that can be caused by using a filtered modulator output as a feedback. Therefore, other embodiments of the invention will be described below, which enhance the loop of FIG. 13 to illustrate this limitation. Printed by the Employees' Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs. Figure 14 is a block diagram of a baseband modulator circuit 1400 designed according to yet another more specific embodiment of the invention. The modulator 1402, the switching stage 1 4 0 4 and the output filter 1 4 0 6 actually operate in a manner similar to the corresponding loop element described above with reference to FIG. 12. According to the corresponding component of the circuit 1 3 0 0, the feedback is provided from the output of the switching stage 1 4 0 4 to the modulator 1 4 0 2 via a divider network including resistors 1 4 1 2 and 1 4 1 4 Feedback input. The feedback is also provided to the feedback input of the modulator 1 402 from its logic output via the filter capacitor 14 0 8 and the series resistor 1 4 1 0. Capacitor 1 408 is selected to attenuate the content of the logic output signal in the useful baseband range and pass the higher frequencies to the feedback path to combine with the attenuated switching level signal. An additional capacitor 1416 is provided from the output of the switching stage 1 4 0 4 to node 1, which will cancel the pole of the feedback path of the capacitor 1408 leading into the switching stage. By looking at the logic output of the modulator 140 2 as ground again and analyzing the feedback path from the output of the switching stage 1 4 04, this point can be understood from the network analysis perspective. If the ratio of the resistors 1 4 1 2 and 1 4 1 4 is in accordance with the Chinese paper standard (CNS) A4 (210X297 mm) -36- 502494 A7 B7 V. Description of the invention (34) Example equal to capacitor 1 4 1 The ratio of 6 to 1 408, then the attenuation of the switching stage output seen at node] _ is the same as that seen at node 2, that is, the voltage levels should be the same. Since the voltages at nodes 1 and 2 are equal, no current flows through resistor 1 4 1 0. Therefore, resistor 1 4 1 0 can be considered as an open circuit, and capacitor 1 4 can be effectively removed from the switching stage feedback path. 0 8 and its corresponding pole. This will make the modulator circuit quite stable and allow delay. Figure 15 is a block diagram of a modulator circuit 1500 designed in accordance with yet a more specific embodiment of the invention. The modulator 1 502, the switching stage 1504, and the output filter 1 5 0 6 actually operate in a manner similar to the corresponding loop element described above with reference to FIG. 12. The resistors 1510, 1512 and 1514 and the capacitors 1508 and 1516 actually operate in a manner similar to the corresponding loop elements described above with reference to FIG. 14. In fact, the embodiment of FIG. 15 includes all the circuits of FIG. 14 plus the feedback path from the output filter 15 0 6 to the modulator stage 1 50 2. The added feedback path is a resistor 1 5 2 0 implementation. Capacitor 1 5 2 2 also contains the poles in the new feedback path introduced from the output of filter 1 5 0 6 to node 1 capacitor 1522, using cancellation capacitors 1508 and 1516. As described in the above embodiment with reference to FIG. 14, if the ratio of the resistor 15 2 0 to the resistor 1 5 1 2 of the parallel resistor 1 5 1 2 is the same as the capacitor 1 5 1 6 of the parallel capacitor 1 5 0 The ratio of 8 to the capacitor 1 5 2 2 means that no current flows into the resistor 1 5 1 0. Therefore, the capacitor 1 5 0 8 ° can be effectively removed from any feedback path. This paper standard applies to China National Standard (CNS) A4 Specifications (210X297 mm) (Please read the notes on the back before filling this page) Order printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs-37- 502494 A7 B7 V. Description of the invention (milk) Please fill in this page again for details) Fig. 16 is a block diagram of a general embodiment of a modulator circuit 1660 designed according to the invention. It replaces the feedback filter and signal mixing circuit of the previously described embodiment with a 4-port network 1 6 0 8, a modulator stage 1 6 0 2, a switching stage 1 6 0 4, and an output filter stage 1 6 0 6 will connect to the 4-port network 1 6 0 8. The network 1608 will obtain the logic output signal of the modulator and attenuate its frequency content within the useful range of the modulator circuit 1600. It also combines the resulting signal with the switching and feedback output of the filter stage to serve as the feedback of the modulator stage 160. Depending on the technology used to implement the modulator circuit 160, the network 16 can be implemented in different ways. In addition, for modulators with amplitude-dependent stability (for example, oversampling and pulse width modulators), it may be necessary to control the amount of modulator feedback related to power and / or filter stage feedback To maintain the best stability point. Therefore, an optional gain control 16 10 is provided before the feedback filter (including the 4-port network 16 0 8) to allow the logic output signal of the modulator to match the gain of the power output. According to a specific embodiment, a digital-to-analog converter (DAC) with variable output amplitude is used for gain control 161 0. Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. Although the invention has been shown and explained with reference to specific embodiments of the invention, those skilled in the art should understand that the disclosed implementation does not depart from the spirit and scope of the invention. The form and details of the examples can be changed. For example, the specific embodiments described above with reference to FIGS. 13 and 14 show the use of analog components (for example, resistors and capacitors) at the output of filters and modulators, combined with the resulting filtered feedback signals and signals from The feedback path of the output of the switching stage. However, it should be understood that, for example, the use of several paper sizes at the modulator level applies to the Chinese National Standard (CNS) A4 specification (210X297 mm) -38- 502494 A7 B7 V. Description of the invention (36) (please first Read the notes on the back and fill in this page) In the case of digital technology implementation, the circuits and technologies used to filter and combine feedback signals can be digital. Similarly, feedback can be implemented through mixed signal modulators, mixed signal circuits, and technologies, such as the signal of the actual ground terminal of a hybrid operational amplifier. It should also be understood that whether the degree of change in the attenuation of the components of the modulator output in the range is appropriate depends on the content of the modulator and the degree of power output in the useful frequency range of the modulator loop, and the tolerable output of the loop Degree of deterioration. That is, it is not necessary to completely reject the frequency components in the output of the modulator stage in the useful loop frequency range in order to maintain reasonable fidelity and stay within the scope of the invention. Printed by the Consumer Cooperative of Intellectual Property Bureau of the Ministry of Economic Affairs. Figure 17 shows the RF bandpass noise spectrum frequency shaping amplifier designed according to the present invention and the technology disclosed in US Patent No. 5,7 7 7,5 1 2. 7 0 0. By way of example, this amplifier can be used in, for example, radiotelephones, pagers, network devices, and so on. The r F amplifier 1700 includes a frequency selective network 1702. The frequency selective network 1702 uses continuous time feedback to apply the modulated R F input to a noise spectral frequency integer. According to a particular embodiment, the network 170 includes at least one resonator stage having a transfer function designed to pass through a frequency band centered at 900 MHz. Of course, it should be understood that the bandpass frequency range of the noise spectral frequency integer amplifier designed according to the invention will range from 1. Any one of a plurality of frequencies such as 8 G Η ζ, 2.4 GHz, etc. is centered, and the present invention is not limited to any specific range or range group. The A / D converter 1 7 0 4 will use the sampling frequency fs to convert the RF signal that has undergone noise spectrum shaping into digital data. According to the specific implementation, this paper standard applies the Chinese National Standard (CNS) A4 specification (210X297 mm) -39-Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 502494 A7 B7 V. Description of the Invention (37) For example, the sampling frequency fs is 3 · 6 G Η z. According to an embodiment, the A / D converter 170 4 includes a comparator. The gate drive circuit 17 0 6 will take the pulse train from the A / D converter 17 0 4 and generate a power output stage for each of the amplifier 1 7 0 — gates of FET 1 7 0 8 and 1 7 1 0 Extremely driven. The output power level shown includes the electric idlers LI, L2, and L3, and the capacitor C1. When the corresponding one of the FETs 1708 and 1710 is closed, this configuration generates two separate resonances at nodes A and B, respectively. Through the feedback path 1 7 1 2, continuous time feedback is provided to the frequency selection network 1 7 2. The output signal of the power level will be sent to the matching network 1 7 1 4 and the matching network 1 7 1 4 will send the output RF signal to the antenna 1 7 1 6 for transmission. As will be understood and in accordance with the present invention, the feedback path 1 7 1 2 may originate within the matching network 1 7 1 4. The loop type of the amplifier 1700 is typically optimally operated with a loop delay equal to 1 / f s (ie, 1/3 · 6GHz). In order to compensate for any increase and unwanted delay related to the output power level, the amplifier 17 0 0 includes another feedback path 1718 to the frequency selective network 1 7 2. The frequency selective network 1702 includes The notch filter at the same center point 1 7 2 0. The source of the feedback is shown as the output of the A / D converter 704. However, it will be understood that there are several other points between this node and the input to the power stage from the feedback source. Therefore, the scope of the invention should not be limited to the embodiments shown. Rather, as with all embodiments described herein, the scope of the present invention should be construed in terms of the scope of the following patent applications. This paper size applies to Chinese National Standard (CNS) A4 specification (210X297 mm) (please read the precautions on the back before filling this page), 11 -40-502494 A7 B7 V. Description of invention (38) The application explains the present invention, but it is obvious that the invention can be implemented in various contents and other embodiments. For example, the RF amplifier of the present invention can be used in a cellular telephone base station or any other application requiring an efficient RF amplifier. The digital amplifier of the present invention can also be used such as 1, 8GHz or 3. RF transmission at 0 GHz or higher. In these embodiments, by increasing the sampling frequency (f s) according to the above ratio, a higher transmission rate can be achieved. It will also be understood that the use of the amplifier configuration and output switching stage of the present invention is not limited to RF applications. Transistors T1 and T2 can also be some different types of devices, including MESFET, HBT, CMOS, or NMOS, and can be made of different processes and materials (including GaAs, SiGe, or standard silicon). Therefore, it can be understood that the embodiments provided herein are merely examples, and the true scope and spirit of the invention should be determined by the scope of patent application. (Please read the notes on the back before filling out this page) Order Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs This paper applies the Chinese National Standard (CNS) A4 specification (210X297 mm) 41-