200302934 玖、發明說明 【相關申請案交互參照】 本申請案係相關於20〇2年2月8日提出申請之審查中 的美國專利申請案，其檔案編號爲OPT-004，此處以引用 的方式將其倂入本文中。 【發明所屬之技術領域】 本發明係關於一種光學資料傳輸器。明確地說，本發 明係關於相較於習知的NRZ光纖光學傳輸器更能夠容忍光 纖色散與非線性現象之光學資料傳輸器。 【先前技術】 在資訊的年代中，對於低成本高資料容量之資料網路 的需求與日倶增。造成此項需求的原因甚多，例如網際網 路與全球資訊網的快速成長便是其中一項原因。網際網路 與全球資訊網中線上使用者的數量不斷地增加，使得因爲 需要極大量頻寬之應用（例如語音與視訊資料流以及檔案傳 送）的不斷增加而大幅地提高對於頻寬的需求。 在提高電信網路頻寬的領域中，光纖傳輸扮演著非常 關鍵的角色。與銅纜比較起來，光纖可提供更高的頻寬， 而且比較不會受到各種電磁干擾及其它不良效應的影響。 因此，對於高速資料率與長距離的資料傳輸而言，其係一 種較佳的媒體。 在非常高速的資料率中，光纖傳輸線中的色散現象會 200302934 導致波形惡化，因而變成標準單模光纖中的一項限制因素 。雖然有色散位移光纖可以使用，而且該種光纖在光纖傳 輸波長中的色散情況非常的低；不過卻已經安裝了大量的 標準單模光纖。因此，非常需要有容忍色散之光學資料傳 輸系統。 可以使用關聯性編碼技術來增強對光纖色散與其它非 線性效應的容忍度。關聯性編碼技術（亦稱爲部份響應信號 處理）係於I960年代所發展出來的。其中一種關聯性編碼 技術稱爲雙二進制信號處理。雙二進制編碼處理最先係由 A.Lender 於 1963 年在 1963 年 5 月的 IEEE Trans· Commun. Electron，vol. CE-82第214-218頁中所發表的文章，名稱 爲用於高速資料傳輸之雙二進制技術（Duobinary Technique for High Speed Data Transmission) 〇 雙二進制（DB)信號的產生方式係藉由將一二進制位元 序列延遲一個完整的位元，然後於原始的位元序列中加入 該延遲二進制位元序列。舉例來說，可參看Franck等人所 提出的美國專利案第5,917,638號。該DB信號可以下面的 方式來表示： (1) 該DB信號係一三位準序列，其頻寬爲該二進制位元 序列m之頻寬的一半。雙二進制編碼處理可將欲進行傳輸 之具有兩個位準的二進制資料信號映對至具有三個意義値 或位準的三位準信號，以縮減該信號的頻寬。舉例來說， 可參看Price等人所提出的美國專利案第5,867,534號。被 200302934 該接收器所接收到的信號則係利用三個位準進行解譯’而 非以兩個位準進行解譯。縮減該信號的頻寬之後便可減少 因色散現象所導致的波形惡化結果。 已經有人利用光學信號來實現雙二進制編碼處理’其 利用的係一偏壓於正交點處的馬赫-曾德爾（Mach-Zehnder) 干涉調變器，並且以一三位準強度的檢波器作爲接收器。 舉例來說，可參看 X.Gu 及 L.C.Blank 於 Electronics Letters Vol.29 Νο·25第2209-2210頁中所發表的文章’名稱爲100 公里標準光纖之1 〇GB/s未重複三位準光學傳輸（1 〇GB/s unrepeatered three-level optical transmission over 100km of standard :fibre)(收錄於 1993 年 10 月 8 日）。 已經有人提出使用二位準（開、關）方式的光學雙二進制 傳輸系統。舉例來說，可參看K.Yonenaga、S.Kuwano、 S.Norimatsu 及 N.Shibata 於 Electronics Letters Vol.31 Νο·4 第302-304頁中所發表的文章，名稱爲無接收器靈敏度降 級之光學雙二進制傳輸系統（《Optical duobiiiary transmission system with no receiver sensitivity degradation)(收錄於1994年12月7日）。因爲典型的光學 檢波器可響應於與振幅相對的光強度，所以可在該檢波器 處自動完成解碼，並且不需要進行雙二進制解碼。該系統 規定處於「開」狀態之信號的相位的値必須爲「〇」或「7Γ 」。該等兩個「開」狀態對應的係該雙二進制信號中的「 + 1」與「-1」狀態，而該「關」狀態對應的則係該雙二進 制信號中的「〇」狀態。 200302934 該光學雙二進制信號的產生方式係藉由驅動一具有推 拉操作之雙驅動的馬赫-曾德爾調變器。其可利用兩個雙二 進制編碼器從原來的二進制信號中產生出用以驅動該馬赫-曾德爾調變器的兩個雙二進制信號。該等兩個雙二進制信 號則會被送至該馬赫-曾德爾調變器的兩個電極中。該雙二 進制信號中的「〇」狀態等於零位準；而對於推拉操作而言 ，「+1」與「-1」狀態的大小相同，但是算術符號相反。 【發明內容】 本發明的容忍色散之光學資料傳輸器會實施前置編碼 。前置編碼可於該線路速率下達成，或於較低的速率下達 成（如果有使用多工器的話）。接收器可利用平方律檢波器來 實施解碼。於其中一實施例中，可以採用小於一個完整位 元週期的延遲。 相較於習知的光學傳輸器，本發明的容忍色散之光學 資料傳輸器受到色散影響的程度約爲四分之一。而且本發 明的容忍色散之光學資料傳輸器亦比較不會受到光纖非線 性現象的影響，並且可以較高的功率位準進行傳輸，由於 抑制了載波的大小，所以可以傳輸更長的距離。 因此，於其中一項觀點中，本發明係具體實現一種包 括一前置編碼器的光學資料傳輸器中，其中該前置編碼器 可將一輸入資料信號分別轉換成輸出位置處的二進制前置 編碼資料信號以及互補輸出位置處的互補二進制前置編碼 資料信號。於其中一實施例中，該前置編碼器係一串列前 200302934 置編碼器。 於另一實施例中，該前置編碼器係一平行前置編碼器 ，其具有η組平行資料輸入用以接收η組平行資料。該平 行前置編碼器會利用該等η組平行資料於η組平行輸出處 產生η組平行前置編碼資料。該平行前置編碼器亦包括一 多工器，其具有分別被耦合至該平行前置編碼器之該等η 組平行輸出的η組平行資料輸入。該多工器會分別產生輸 出位置處的二進制前置編碼資料信號以及互補輸出位置處 _ 的互補二進制前置編碼資料信號。 該光學資料傳輸器亦包括一被耦合至該前置編碼器之 輸出與互補輸出中其中一者的延遲元件。該延遲元件會藉 由將該互補二進制前置編碼資料信號與該二進制前置編碼 資料信號中其中一者相對於另一者延遲一段時間，以便於 該延遲元件輸出處產生一延遲資料信號，該段時間小於該 二進制前置編碼資料信號的一個位元週期。 於其中一實施例中，該延遲元件將該互補二進制前置 φ 編碼資料信號與該二進制前置編碼資料信號中其中一者相 對於另一者延遲一段時間，該時間係介於該二進制前置編 碼資料信號之位元週期的0.4倍至0.8倍之間。於其中一實 施例中，該延遲元件包括一可變的延遲元件。於其中一實 施例中’該延遲兀件經過選擇之後可增加包3旨亥光學貝料 傳輸器之通信系統的色散容忍能力。 該光學資料傳輸器亦包括一光學資料調變器’其具有 一被耦合至該前置編碼器之輸出與互補輸出中其中一者的 11 200302934 第一資料輸入。該光學資料調變器包括一被耦合至該延遲 元件之輸出的第二資料輸入。該光學資料調變器係響應該 延遲資料信號以及該二進制前置編碼資料信號與該互補二 進制前置編碼資料信號中其中一者，以便對被送至該光學 資料調變器之光學輸入處的光學信號進行調變，以產生一 調變後的光學輸出信號。 於其中一實施例中，該光學資料調變器包括一差動輸 入的馬赫-曾德爾調變器。於其中一實施例中，該光學資料 φ 調變器經過偏壓之後，當該延遲資料信號以及被耦合至該 光學資料調變器之第一資料輸入處之該二進制前置編碼資 料信號與該互補二進制前置編碼資料信號中其中一者之間 的差異等於該差異最大値與該差異最小値之平均値時，該 經過調變之光學輸出信號便會具有最小的強度。 於其中一實施例中，該光學資料傳輸器亦包括一偏壓 電壓源，其係調整該二進制前置編碼資料信號、該互補二 進制前置編碼資料信號以及該延遲資料信號中至少其中一 φ 者的平均振幅。 於另一項觀點中，本發明係具體實現於一種含有一前 置編碼器之光學資料傳輸器中，該前置編碼器係將一輸入 資料信號轉換成輸出處的二進制前置編碼資料信號。於其 中一實施例中，該前置編碼器係一串列前置編碼器。 於另一實施例中，該前置編碼器係一平行前置編碼器 ，其具有η組平行資料輸入用以接收η組平行資料。該平 行前置編碼器會利用該等η組平行資料於η組平行輸出處 12 200302934 產生η組平行前置編碼資料。該平行前置編碼器亦包括一 多工器，其具有分別被耦合至該平行前置編碼器之該等η 組平行輸出的η組平行資料輸入。該多工器會在輸出位置 處產生該二進制前置編碼資料信號。 該光學資料傳輸器亦包括一放大器，其具有一被耦合 至該前置編碼器之輸出的輸入。該放大器會將該二進制前 置編碼資料信號分別放大成輸出位置處的放大資料信號以 及互補輸出位置處的互補放大資料信號。 該光學資料傳輸器亦包括一被耦合至該放大器之輸出 與互補輸出中其中一者的延遲元件。該延遲元件會藉由將 該放大資料信號與該互補放大資料信號中其中一者相對於 另一者延遲一段時間以便於該延遲元件之輸出位置處產生 一延遲放大資料信號，該段延遲時間小於該二進制前置編 碼資料信號的一個位元週期。 於其中一實施例中，該延遲元件會將該放大資料信號 與該互補放大資料信號中其中一者相對於另一者延遲一段 時間，該時間係介於該二進制前置編碼資料信號之位元週 期的0.4倍至0.8倍之間。於其中一實施例中，該延遲元件 包括一可變的延遲元件。於其中一實施例中，該延遲元件 經過選擇之後可增加包含該光學資料傳輸器之通信系統的 色散容忍能力。 該光學資料傳輸器進一步包括一光學資料調變器，其 具有一被耦合至該放大器之輸出與互補輸出中其中一者的 第一資料輸入，以及具有一被耦合至該延遲元件之輸出的 13 200302934 第二資料輸入。該光學資料調變器可響應該延遲放大資料 信號以及該放大資料信號與該互補放大資料信號中其中一 者，以便對被送至該光學資料調變器之光學輸入處的光學 信號進行調變，以產生一調變後的光學輸出信號。 於其中一實施例中，該光學資料調變器包括一差動輸 入的馬赫-曾德爾調變器。於其中一實施例中，該光學資料 調變器經過偏壓之後，當該延遲放大資料信號以及被耦合 至該光學資料調變器之第一資料輸入處之該放大資料信號 與該互補放大資料信號中其中一者之間的差異等於該差異 最大値與該差異最小値之平均値時，該經過調變之光學輸 出信號便會具有最小的強度。 於其中一實施例中，該光學資料傳輸器亦包括一偏壓 電壓源，其係調整該放大資料信號、該互補放大資料信號 以及該延遲放大資料信號中至少其中一者的平均振幅。 於另一項觀點中，本發明係具體實現於一種用於編碼 一光學資料信號的方法中。該方法包括將一輸入資料信號 轉換成一二進制前置編碼資料信號以及一互補二進制前置 編碼資料信號。於其中一實施例中，該輸入資料信號係一 串列資料信號。於另一實施例中，將該輸入資料信號轉換 成一二進制前置編碼資料信號以及一互補二進制前置編碼 資料信號包括將η組平行資料信號轉換成η組平行前置編 碼資料信號，並且對該等η組平行前置編碼資料信號進行 多工處理，以產生該二進制前置編碼資料信號以及該互補 二進制前置編碼資料信號。 14 200302934 該方法亦包括藉由將該二進制前置編碼資料信號與該 互補二進制前置編碼資料信號中其中一者相對於另一者延 遲一段小於該二進制前置編碼資料信號的一個位元週期的 時間，以產生一延遲資料信號。於其中一實施例中，產生 該延遲資料信號包括將該二進制前置編碼資料信號與該互 補二進制前置編碼資料信號中其中一者相對於另一者延遲 一段介於該二進制前置編碼資料信號之位元週期的0.4倍 至0.8倍之間的時間。於其中一實施例中，產生該延遲資 料信號包括延遲一段時間，以便使用該方法來編碼一光學 資料信號而提高通信系統的色散容忍能力。 該方法進一步包括響應該延遲資料信號以及該二進制 前置編碼資料信號與該互補二進制前置編碼資料信號中其 中一者，以調變一光學信號，以便產生一經過調變之光學 輸出信號。於其中一實施例中，該調變進一步包括偏壓一 調變器，使得當該延遲資料信號以及該二進制前置編碼資 料信號與該互補二進制前置編碼資料信號中其中一者之間 的差異等於該差異最大値與該差異最小値之平均値時，該 經過調變之光學輸出信號的光學強度便會具有最小値。於 其中一實施例中，該方法進一步包括調整該二進制前置編 碼資料信號、該互補二進制前置編碼資料信號以及該延遲 資料信號中至少其中一者的平均振幅。 於另一項觀點中，本發明係具體實現一種甩·於編碼一 光學資料信號的方法。該方法包括將一輸入資料信號轉換 成一二進制前置編碼資料信號。於其中一實施例中，該輸 15 200302934 入資料信號係一串列資料信號。於另一實施例中，將該輸 入資料信號轉換成一二進制前置編碼資料信號包括將η組 平行資料信號轉換成η組平行前置編碼資料信號，並且對 該等η組平行前置編碼資料信號進行多工處理，以產生該 二進制前置編碼資料信號。 該方法進一步包括放大該二進制前置編碼資料信號， 以產生一放大資料信號以及一互補放大資料信號。 該方法亦包括藉由將該放大資料信號以及該互補放大 資料信號中其中一者相對於另一者延遲一段小於該二進制 前置編碼資料信號的一個位元週期的時間，以產生一延遲 放大資料信號。 於其中一實施例中，產生該延遲放大資料信號包括將 該放大資料信號以及該互補放大資料信號中其中一者相對 於另一者延遲一段介於該二進制前置編碼資料信號之位元 週期的〇_4倍至0.8倍之間的時間。於其中一實施例中，產 生該延遲放大資料信號包括延遲一段時間，以便使用該方 法來編碼一光學資料信號而提高通信系統的色散容忍能力 〇 該方法進一步包括響應該延遲放大資料信號以及該放 大資料信號與該互補放大資料信號中其中一者，以調變一 光學信號，以便產生一經過調變之光學輸出信號。於其中 一實施例中，該調變進一步包括偏壓一調變器，使得當該 延遲放大資料信號以及該放大資料信號與該互補放大資料 信號中其中一者之間的差異等於該差異最大値與該差異最 200302934 小値之平均値時，該經過調變之光學輸出信號的光學強度 便會具有最小値。於其中一實施例中，該方法進一步包括 調整該放大資料信號、該互補放大資料信號以及該延遲放 大資料信號中至少其中一者的平均振幅。 【實施方式】 圖1顯示的係根據本發明之容忍色散之傳輸器100的 一具體實施例，其包括一串列資料輸入104。該傳輸器100 包括一串列雙二進制前置編碼器102。該串列雙二進制前置 編碼器102具有一串列資料輸入104，其可接收一輸入資料 信號。該雙二進制前置編碼器102亦具有一輸出106以及 一互補輸出108，用以分別產生一二進制前置編碼資料信號 以及一互補二進制前置編碼資料信號。 該傳輸器100亦包括一延遲元件110，其具有一被耦合 至該雙二進制前置編碼器102之互補輸出108的輸入112。 於其中一實施例中，該延遲元件110會將該互補二進制前 置編碼資料信號相對於該二進制前置編碼資料信號延遲一 段小於該二進制前置編碼資料信號的一個位元週期的時間 0 該傳輸器100亦包括一差動放大器114，其具有一被耦 合至該延遲元件110之輸出120的第一輸入118，以及一被 耦合至該前置編碼器102之輸出106的第二輸入122。於另 一實施例中（圖中未顯示），該延遲元件110係被耦合於該雙 二進制前置編碼器102之輸出106與該差動放大器114之 17 200302934 第二輸入122之間。於此實施例中，該雙二進制前置編碼 器102之互補輸出108則係被直接耦合至該差動放大器114 之第一輸入118。 於另一實施例中（圖中未顯示），該延遲元件110係被耦 合於該雙二進制前置編碼器102之互補輸出108與該差動 放大器114之第一輸入118之間。第二延遲元件（圖中未顯 示）則係被耦合於該雙二進制前置編碼器102之輸出106與 該差動放大器Π4之第二輸入122之間。 於其中一實施例中，則可將該延遲元件110整合至一 條纜線（圖中未顯示）中，該條纜線可將該雙二進制前置編碼 器102之互補輸出108耦合至該差動放大器114之第一輸 入118。於另一實施例中，該延遲元件110則包括該條纜線 本身，透過該條纜線來傳播該互補二進制前置編碼資料信 號便可產生該延遲。於此實施例中，該條纜線的長度必須 經過選擇以對應該延遲。 該差動放大器Π4可將該二進制前置編碼資料信號以 及該互補二進制前置編碼資料信號分別轉換成差動輸出124 之上的差動信號g⑴以及互補差動輸出126之上的互補差 動信號-g⑴。該差動信號g⑴以及該互補差動信號-g⑴都係 四位準資料信號。四位準資料信號意謂著該差動信號g⑴ 以及該互補差動信號-g⑴皆包括四個（4個）有意義的數値或 位準。 於其中一實施例中，存在一偏壓網路（例如T型偏壓源 128)，用以增加一偏壓電壓給該差動信號g⑴以及該互補差 18 200302934 動信號-g⑴中其中一者。於圖中所示的實施例中，該τ型 偏壓源128係增加一偏壓電壓給該互補差動信號_g⑴，用 以產生一含有DC偏移電壓之互補差動信號-g(t)+Vbias。於 另一實施例中，則係由該差動放大器114產生一含有偏壓 電壓之互補差動信號-g(t)。 此外，該傳輸器100包括一光學資料調變器130。本發 明之傳輸器100可以使用各種類型的光學資料調變器。於 圖中所示的實施例中，該光學資料調變器130係一差動輸 入的馬赫-曾德爾干涉（MZI)資料調變器，其具有分別被耦 合至該差動放大器114之差動輸出124與互補差動輸出126 的第一資料輸入132與第二資料輸入134。 差動輸入MZI資料調變器的優點非常多，因爲與單輸 入的MZI資料調變器比較起來，其僅需要較低的驅動電壓 。該差動輸入資料調變器130每個輸入所需要的驅動電壓 僅爲一單輸入資料調變器（圖中未顯示）所需要的驅動電壓的 一半。 於圖中所示的實施例中，該差動放大器114係以推拉 的方式直接驅動該差動輸入資料調變器130。因此，於此實 施例中，該差動放大器114既是該差動輸入資料調變器130 的驅動器；同時也是一放大器，用以將該二進制前置編碼 資料信號以及該互補二進制前置編碼資料信號分別轉換成 四位準資料信號以及互補四位準資料信號。 於其中一實施例中，該光學資料傳輸器100進一步包 括一濾波器13 5，其會被親合至該差動放大器114之差動輸 19 200302934 出124與互補差動輸出126中其中一者，並且會被耦合至 該光學資料調變器130之第一資料輸入132與第二資料輸 入134中其中一者。該濾波器135可縮減該四位準資料信 號與該互補四位準資料信號中至少其中一者的頻寬。於其 中一實施例中，該濾波器135可提供一可調整之截止頻率 〇 於另一實施例中，該光學資料傳輸器1〇〇進一步包括 一第一濾波器1 3 5與一第二濾波器13 7。該第一濾波器1 3 5 會被耦合至該差動放大器114之差動輸出124，以及被耦合 至該光學資料調變器130之第一資料輸入132。該第二濾波 器137則會被耦合至該差動放大器114之互補差動輸出126 ，以及被耦合至該光學資料調變器130之第二資料輸入134 。該等第一濾波器135與第二濾波器137可分別縮減該四 位準資料信號與該互補四位準資料信號的頻寬。於其中一 實施例中，該等第一濾波器135與第二濾波器137中至少 其中一者可提供一可調整之截止頻率。 該資料調變器130亦包括一光學輸入136，用以從一光 源（例如雷射138)接收一光學信號。該資料調變器130可響 應該差動信號g⑴以及該含有DC偏移電壓之互補差動信號 -g⑴+Vbias對一連續波光學信號的振幅進行調變，並且產生 一經過調變的光學輸出信號。 與習知的NRZ光纖傳輸器及先前技術中的雙二進制傳 輸器比較起來，該容忍色散之傳輸器1〇〇對於光纖色散效 應的容忍能力比較強，其至少部分原因係因爲該延遲元件 20 200302934 110所產生的延遲小於一個完整的位元週期而使得該經過調 變的光學信號之光學振幅會以零値爲中心進行震盪。當該 信號沿著該條光纖進行傳播時，該等震盪結果會趨於衰減 成零値，因而便可保留1位準的振幅，而且不會損及零位 準的信號品質。該等震盪的峰値係出現在該等位元的邊界 處，因此該等震盪並不會損及消失率（extinction rate)。在 先前技術具有完整位元週期之延遲時間的雙二進制信號中 ，則不會出現該些震盪情形。 於另一實施例中，該光學資料調變器130爲一單輸入 零響聲的馬赫-曾德爾調變器（圖中未顯示）。於此實施例中 ，該零響聲的MZI調變器包括一被耦合至該差動放大器 114之差動輸出124與互補差動輸出126中其中一者的單一 輸入。該零響聲的MZI調變器可由該差動信號g⑴以及該 互補差動信號-g⑴中其中一者來驅動。 與操作於相同位元率的光學資料調變器比較起來，該 光學資料調變器130的頻寬比較低，其係因爲用以驅動該 調變器130的四位準資料信號具有較低頻寬的關係。因此 ，本發明的傳輸器1〇〇可以使用較廉價的光學資料調變器 ，從而便可降低該光學資料傳輸器1〇〇總體的成本。 於其中一實施例中，該光學資料調變器130之前置設 定偏壓點的設計可使其不需要使用一分離的偏壓電壓。於 另一實施例中，該光學資料調變器130包括一分離的輸入( 圖中未顯示）用以施加一偏壓電壓給該調變器130。於此實 施例中，會施加一分離的偏壓電壓源（圖中未顯示）給該輸入 21 200302934 ，以便設定該調變器的操作點。 於操作中，該光學傳輸器100會在該雙二進制前置編 碼器102的串列資料輸入104處，以1/T的位元率接收一 連串的二進制資料位元（d)流，其中T爲該位元週期。該雙 二進制前置編碼器102會對該等二進制資料位元（d)進行編 碼。該雙二進制前置編碼器102則會在該雙二進制前置編 碼器102的輸出106處及互補輸出108處分別產生一二進 制前置編碼位元序列m及一互補二進制前置編碼位元序列 5。於其中一實施例中，可根據下面的公式產生該二進制前 置編碼位兀序列m : mfmM ㊉ & (2) 其中i爲位元指標，而㊉爲邏輯互斥或（XOR)運算元。 時間變動電氣信號m⑴及；^t)則分別對應該二進制前置編 碼位元序列m及該互補二進制前置編碼位元序列叾。 該延遲元件110會將該互補前置編碼資料信號相 對於該前置編碼資料信號m⑴於時間上產生延遲。於其中 一實施例中，該延遲元件110會將該互補前置編碼資料信 號叾⑴相對於該前置編碼資料信號m⑴於時間上實質產生 介於〇·4Τ至0.9T之間的延遲長度。最佳的時間延遲則係 取決於該等信號的頻寬以及該差動放大器114的線性特性 。舉例來說，當頻寬爲位元率之百'分之七十五（75%)而且是 一線性差動放大器時，最佳的時間延遲約爲0.75Τ。 該差動放大器114會接收該前置編碼資料信號m(t)以 及該延遲互補前置編碼資料信號5(t)，並且將該二進制前 22 200302934 置編碼資料信號m⑴以及該互補二進制前置編碼資料信號 叾⑴分別轉換成該差動輸出124之上的四位準資料信號g⑴ 以及該互補差動輸出126之上的互補四位準資料信號-g⑴ 。該四位準資料信號g⑴的表示如下： g⑴=k(m⑴-5(t- τ )) (3)200302934 发明 Description of the invention [Cross-reference to related applications] This application is related to the pending US patent application filed on February 8, 2002. Its file number is OPT-004, which is hereby incorporated by reference. Incorporate it into this article. [Technical field to which the invention belongs] The present invention relates to an optical data transmitter. Specifically, the present invention relates to an optical data transmitter that is more tolerant to optical fiber dispersion and nonlinear phenomena than the conventional NRZ fiber optical transmitter. [Previous technology] In the age of information, the demand for data networks with low cost and high data capacity is increasing. There are many reasons for this demand, such as the rapid growth of the Internet and the World Wide Web. The continuous increase in the number of online users on the Internet and the World Wide Web has led to a significant increase in demand for bandwidth due to the increasing number of applications that require a large amount of bandwidth, such as voice and video data streams and file transfers. In the field of increasing the bandwidth of telecommunication networks, fiber optic transmission plays a key role. Compared with copper cables, optical fibers can provide higher bandwidth and are less affected by various electromagnetic interference and other adverse effects. Therefore, it is a better medium for high-speed data rate and long-distance data transmission. At very high data rates, dispersion in optical fiber transmission lines can cause 200302934 to worsen the waveform and thus become a limiting factor in standard single-mode fibers. Although dispersion-shifted fiber can be used, and the dispersion of this fiber in the transmission wavelength of the fiber is very low; however, a large number of standard single-mode fibers have been installed. Therefore, there is a great need for optical data transmission systems that tolerate dispersion. Correlation coding techniques can be used to enhance tolerance to fiber dispersion and other non-linear effects. Correlation coding technology (also known as partial response signal processing) was developed in the I960s. One such correlation coding technique is called dual binary signal processing. The bi-binary encoding process was first published by A. Lender in 1963, IEEE Trans · Commun. Electron, vol. CE-82, pp. 214-218, May 1963. It was named for high-speed data transmission. Duobinary Technique for High Speed Data Transmission 〇 Binary binary (DB) signal is generated by delaying a binary bit sequence by a complete bit, and then adding the delay to the original bit sequence Binary bit sequence. See, for example, U.S. Patent No. 5,917,638 filed by Franck et al. The DB signal can be expressed in the following manner: (1) The DB signal is a three-bit quasi-sequence whose bandwidth is half the bandwidth of the binary bit sequence m. The dual binary coding process can map a binary data signal with two levels to be transmitted to a three-level signal with three meanings or levels to reduce the bandwidth of the signal. For example, see US Patent No. 5,867,534 filed by Price et al. The signal received by the 200302934 receiver was interpreted using three levels' instead of interpreting at two levels. Reducing the bandwidth of this signal can reduce the waveform deterioration caused by dispersion. Some people have used optical signals to implement bi-binary coding processing. Its use is a Mach-Zehnder interference modulator biased at orthogonal points, and a three-level quasi-intensity detector is used receiver. For example, see X.Gu and LCBlank's article in Electronics Letters Vol. 29 No. 25, 2209-2210, entitled '100GB / s Unreplicated Tri-level Optics for 100km Standard Fiber Transmission (100GB / s unrepeatered three-level optical transmission over 100km of standard: fibre) (incorporated on October 8, 1993). Optical double-binary transmission systems using two-level (on, off) methods have been proposed. For example, see K.Yonenaga, S.Kuwano, S.Norimatsu, and N.Shibata in Electronics Letters Vol. 31 No. 4 pages 302-304, titled Optics without Degradation of Receiver Sensitivity Dual binary transmission system ("Optical duobiiiary transmission system with no receiver sensitivity degradation") (contained on December 7, 1994). Because a typical optical detector is responsive to the intensity of the light relative to the amplitude, decoding can be done automatically at the detector and no dual binary decoding is required. This system stipulates that the phase of the signal in the "on" state must be "0" or "7Γ". The two “on” states correspond to the “+1” and “-1” states in the dual binary signal, while the “off” states correspond to the “0” state in the dual binary signal. 200302934 The optical dual binary signal is generated by driving a dual-drive Mach-Zehnder modulator with push-pull operation. It can use two dual binary encoders to generate two dual binary signals from the original binary signals to drive the Mach-Zehnder modulator. The two binary signals are sent to the two electrodes of the Mach-Zehnder modulator. The "0" state in this double binary signal is equal to the zero level; for the push-pull operation, the "+1" and "-1" states have the same size, but the arithmetic sign is opposite. SUMMARY OF THE INVENTION The dispersion-tolerant optical data transmitter of the present invention implements precoding. Precoding can be achieved at this line rate, or at a lower rate (if a multiplexer is used). The receiver can use a square-law detector to perform the decoding. In one embodiment, a delay of less than one full bit period may be used. Compared with the conventional optical transmitter, the dispersion-tolerant optical data transmitter of the present invention is affected by dispersion to about a quarter. Moreover, the dispersion-tolerant optical data transmitter of the present invention is also less affected by the non-linearity of the optical fiber, and can transmit at a higher power level. Because the size of the carrier is suppressed, it can transmit a longer distance. Therefore, in one aspect, the present invention specifically implements an optical data transmitter including a precoder, wherein the precoder can respectively convert an input data signal into a binary preamble at an output position. A coded data signal and a complementary binary pre-coded data signal at a complementary output position. In one embodiment, the pre-encoder is a series of pre-200302934 encoders. In another embodiment, the pre-encoder is a parallel pre-encoder having n parallel data inputs for receiving n parallel data. The parallel pre-encoder will generate n parallel pre-encoded data at the n parallel output using these n parallel data. The parallel pre-encoder also includes a multiplexer having n sets of parallel data inputs respectively coupled to the n sets of parallel outputs of the parallel pre-encoder. The multiplexer generates a binary pre-encoded data signal at the output position and a complementary binary pre-encoded data signal at _ at the complementary output position, respectively. The optical data transmitter also includes a delay element coupled to one of an output of the precoder and a complementary output. The delay element delays one of the complementary binary pre-encoded data signal and the binary pre-encoded data signal relative to the other for a period of time in order to generate a delayed data signal at the output of the delay element. The segment time is shorter than one bit period of the binary pre-encoded data signal. In one embodiment, the delay element delays one of the complementary binary pre-coded φ-coded data signal and the binary pre-coded data signal relative to the other for a period of time, the time being between the binary pre- The bit period of the encoded data signal is between 0.4 and 0.8 times. In one embodiment, the delay element includes a variable delay element. In one of the embodiments, the delay element is selected to increase the dispersion tolerance of the communication system of the optical transmission device. The optical data transmitter also includes an optical data modulator 'having an 11 200302934 first data input coupled to one of an output and a complementary output of the pre-encoder. The optical data modulator includes a second data input coupled to an output of the delay element. The optical data modulator responds to one of the delayed data signal and the binary pre-encoded data signal and the complementary binary pre-encoded data signal, so that an optical input to the optical input of the optical data modulator The optical signal is modulated to generate a modulated optical output signal. In one embodiment, the optical data modulator includes a Mach-Zehnder modulator with a differential input. In one embodiment, after the optical data φ modulator is biased, when the delayed data signal and the binary pre-encoded data signal coupled to the first data input of the optical data modulator and the optical data modulator are biased, When the difference between one of the complementary binary pre-encoded data signals is equal to the average of the maximum difference and the minimum difference, the modulated optical output signal will have the smallest intensity. In one embodiment, the optical data transmitter also includes a bias voltage source that adjusts at least one of the binary precoded data signal, the complementary binary precoded data signal, and the delayed data signal. Average amplitude. In another aspect, the present invention is embodied in an optical data transmitter including a pre-encoder. The pre-encoder converts an input data signal into a binary pre-encoded data signal at the output. In one embodiment, the precoder is a series of precoders. In another embodiment, the pre-encoder is a parallel pre-encoder having n parallel data inputs for receiving n parallel data. The parallel pre-encoder will use the n parallel data at the n parallel output 12 200302934 to generate n parallel pre-encoded data. The parallel pre-encoder also includes a multiplexer having n sets of parallel data inputs respectively coupled to the n sets of parallel outputs of the parallel pre-encoder. The multiplexer generates the binary pre-encoded data signal at the output position. The optical data transmitter also includes an amplifier having an input coupled to an output of the pre-encoder. The amplifier amplifies the binary pre-encoded data signal into an amplified data signal at an output position and a complementary amplified data signal at a complementary output position, respectively. The optical data transmitter also includes a delay element coupled to one of an output of the amplifier and a complementary output. The delay element delays one of the amplified data signal and the complementary amplified data signal with respect to the other for a period of time in order to generate a delayed amplified data signal at the output position of the delay element. One bit period of the binary pre-encoded data signal. In one embodiment, the delay element delays one of the amplified data signal and the complementary amplified data signal relative to the other for a period of time, the time being between the bits of the binary pre-encoded data signal Between 0.4 and 0.8 times the period. In one embodiment, the delay element includes a variable delay element. In one embodiment, the delay element is selected to increase the dispersion tolerance of the communication system including the optical data transmitter. The optical data transmitter further includes an optical data modulator having a first data input coupled to one of an output of the amplifier and a complementary output, and a 13 having an output coupled to the delay element. 200302934 Second data entry. The optical data modulator can respond to the delayed amplified data signal and one of the amplified data signal and the complementary amplified data signal, so as to modulate the optical signal sent to the optical input of the optical data modulator. To generate a modulated optical output signal. In one embodiment, the optical data modulator includes a Mach-Zehnder modulator with a differential input. In one embodiment, after the optical data modulator is biased, when the delayed amplified data signal and the amplified data signal and the complementary amplified data are coupled to the first data input of the optical data modulator, When the difference between one of the signals is equal to the average of the largest difference and the smallest difference, the modulated optical output signal will have the smallest intensity. In one embodiment, the optical data transmitter also includes a bias voltage source that adjusts the average amplitude of at least one of the amplified data signal, the complementary amplified data signal, and the delayed amplified data signal. In another aspect, the invention is embodied in a method for encoding an optical data signal. The method includes converting an input data signal into a binary pre-coded data signal and a complementary binary pre-coded data signal. In one embodiment, the input data signal is a serial data signal. In another embodiment, converting the input data signal into a binary precoded data signal and a complementary binary precoded data signal includes converting n parallel data signals into n parallel parallel coded data signals, and The η groups of parallel pre-encoded data signals are multiplexed to generate the binary pre-encoded data signal and the complementary binary pre-encoded data signal. 14 200302934 The method also includes delaying one of the binary precoded data signal and the complementary binary precoded data signal with respect to the other by a period of less than one bit period of the binary precoded data signal. Time to generate a delayed data signal. In one embodiment, generating the delayed data signal includes delaying one of the binary precoded data signal and the complementary binary precoded data signal relative to the other by a period between the binary precoded data signal The time between 0.4 and 0.8 times the bit period. In one embodiment, generating the delayed data signal includes delaying a period of time, so as to use the method to encode an optical data signal to improve the dispersion tolerance of the communication system. The method further includes responding to the delayed data signal and one of the binary pre-encoded data signal and the complementary binary pre-encoded data signal to modulate an optical signal so as to generate a modulated optical output signal. In one embodiment, the modulation further includes biasing a modulator such that when the delayed data signal and the difference between the binary pre-encoded data signal and one of the complementary binary pre-encoded data signals When equal to the average of the maximum difference and the minimum difference, the optical intensity of the modulated optical output signal will have a minimum value. In one embodiment, the method further includes adjusting an average amplitude of at least one of the binary precoded data signal, the complementary binary precoded data signal, and the delayed data signal. In another aspect, the present invention specifically implements a method for encoding an optical data signal. The method includes converting an input data signal into a binary pre-encoded data signal. In one embodiment, the input data signal is a series of data signals. In another embodiment, converting the input data signal into a binary pre-encoded data signal includes converting n parallel data signals into n parallel pre-coded data signals, and for the n parallel pre-encoded data signals. Multiplexing is performed to generate the binary pre-encoded data signal. The method further includes amplifying the binary pre-encoded data signal to generate an amplified data signal and a complementary amplified data signal. The method also includes generating a delayed amplified data by delaying one of the amplified data signal and the complementary amplified data signal relative to the other for a period of time less than a bit period of the binary pre-encoded data signal. signal. In one embodiment, generating the delayed amplified data signal includes delaying one of the amplified data signal and the complementary amplified data signal relative to the other by a bit period between the binary pre-encoded data signal. 〇_4 times to 0.8 times. In one embodiment, generating the delayed amplified data signal includes delaying a period of time, so as to use the method to encode an optical data signal to improve the dispersion tolerance of the communication system. The method further includes responding to the delayed amplified data signal and the amplification. One of the data signal and the complementary amplified data signal is used to modulate an optical signal so as to generate a modulated optical output signal. In one embodiment, the modulation further includes biasing a modulator so that when the delayed amplified data signal and a difference between the amplified data signal and one of the complementary amplified data signals are equal to the maximum difference 値When the difference is the maximum mean value of 200302934 hours, the optical intensity of the modulated optical output signal will have the smallest value. In one embodiment, the method further includes adjusting an average amplitude of at least one of the amplified data signal, the complementary amplified data signal, and the delayed amplified data signal. [Embodiment] FIG. 1 shows a specific embodiment of a dispersion-tolerant transmitter 100 according to the present invention, which includes a series of data inputs 104. The transmitter 100 includes a series of dual binary precoders 102. The serial bi-binary pre-encoder 102 has a serial data input 104 which can receive an input data signal. The dual binary precoder 102 also has an output 106 and a complementary output 108 for generating a binary precoded data signal and a complementary binary precoded data signal, respectively. The transmitter 100 also includes a delay element 110 having an input 112 coupled to a complementary output 108 of the dual binary pre-encoder 102. In one embodiment, the delay element 110 delays the complementary binary pre-encoded data signal relative to the binary pre-encoded data signal by a time less than one bit period of the binary pre-encoded data signal. 0 The transmission The encoder 100 also includes a differential amplifier 114 having a first input 118 coupled to the output 120 of the delay element 110 and a second input 122 coupled to the output 106 of the pre-encoder 102. In another embodiment (not shown), the delay element 110 is coupled between the output 106 of the dual binary pre-encoder 102 and the second input 122 of 17 200302934 of the differential amplifier 114. In this embodiment, the complementary output 108 of the dual binary pre-encoder 102 is directly coupled to the first input 118 of the differential amplifier 114. In another embodiment (not shown), the delay element 110 is coupled between the complementary output 108 of the dual binary pre-encoder 102 and the first input 118 of the differential amplifier 114. A second delay element (not shown) is coupled between the output 106 of the dual binary pre-encoder 102 and the second input 122 of the differential amplifier Π4. In one embodiment, the delay element 110 can be integrated into a cable (not shown), which can couple the complementary output 108 of the dual binary pre-encoder 102 to the differential First input 118 of the amplifier 114. In another embodiment, the delay element 110 includes the cable itself, and the delay can be generated by propagating the complementary binary pre-encoded data signal through the cable. In this embodiment, the length of the cable must be selected to correspond to the delay. The differential amplifier Π4 can convert the binary pre-encoded data signal and the complementary binary pre-encoded data signal into a differential signal g⑴ above the differential output 124 and a complementary differential signal above the complementary differential output 126, respectively. -g⑴. The differential signal g⑴ and the complementary differential signal -g⑴ are all four-level data signals. The four-level data signal means that the differential signal g⑴ and the complementary differential signal -g⑴ each include four (4) significant numbers or levels. In one embodiment, there is a bias network (such as a T-shaped bias source 128) for adding a bias voltage to the differential signal g⑴ and the complementary difference 18 200302934 dynamic signal -g⑴ . In the embodiment shown in the figure, the τ-type bias source 128 adds a bias voltage to the complementary differential signal _g⑴ to generate a complementary differential signal -g (t ) + Vbias. In another embodiment, the differential amplifier 114 generates a complementary differential signal -g (t) with a bias voltage. In addition, the transmitter 100 includes an optical data modulator 130. The transmitter 100 of the present invention can use various types of optical data modulators. In the embodiment shown in the figure, the optical data modulator 130 is a Mach-Zehnder interference (MZI) data modulator with a differential input, which has differentials respectively coupled to the differential amplifier 114. The first data input 132 and the second data input 134 of the output 124 and the complementary differential output 126. Differential input MZI data modulators have many advantages, because they only require lower drive voltages compared to single-input MZI data modulators. The driving voltage required for each input of the differential input data modulator 130 is only half of the driving voltage required for a single input data modulator (not shown). In the embodiment shown in the figure, the differential amplifier 114 directly drives the differential input data modulator 130 in a push-pull manner. Therefore, in this embodiment, the differential amplifier 114 is not only the driver of the differential input data modulator 130, but also an amplifier for the binary pre-encoded data signal and the complementary binary pre-encoded data signal. Converted into four-level data signals and complementary four-level data signals, respectively. In one embodiment, the optical data transmitter 100 further includes a filter 13 5 which is coupled to one of the differential output 19 200302934 output 124 and the complementary differential output 126 of the differential amplifier 114. And is coupled to one of the first data input 132 and the second data input 134 of the optical data modulator 130. The filter 135 can reduce the bandwidth of at least one of the four-level data signal and the complementary four-level data signal. In one embodiment, the filter 135 can provide an adjustable cut-off frequency. In another embodiment, the optical data transmitter 100 further includes a first filter 135 and a second filter.器 13 7． The first filter 1 3 5 is coupled to a differential output 124 of the differential amplifier 114 and a first data input 132 of the optical data modulator 130. The second filter 137 is coupled to a complementary differential output 126 of the differential amplifier 114 and a second data input 134 of the optical data modulator 130. The first filter 135 and the second filter 137 can reduce the bandwidth of the four-level data signal and the complementary four-level data signal, respectively. In one embodiment, at least one of the first filter 135 and the second filter 137 may provide an adjustable cut-off frequency. The data modulator 130 also includes an optical input 136 for receiving an optical signal from a light source, such as a laser 138. The data modulator 130 can respond to the differential signal g⑴ and the complementary differential signal -g⑴ + Vbias containing a DC offset voltage to modulate the amplitude of a continuous wave optical signal, and generate a modulated optical output signal. Compared with the conventional NRZ optical fiber transmitter and the dual binary transmitter in the prior art, the dispersion-tolerant transmitter 100 has a stronger tolerance for fiber dispersion effects, at least in part because of the delay element 20 200302934 The delay generated by 110 is less than a complete bit period, so that the optical amplitude of the modulated optical signal will oscillate around zero chirp. When the signal propagates along the fiber, the oscillation results tend to decay to zero, so that the amplitude of 1 level can be retained without compromising the signal quality of the zero level. The peaks of these oscillations appear at the boundaries of these bits, so the oscillations do not damage the extinction rate. In the prior art dual binary signals with a delay time of a full bit period, these oscillations would not occur. In another embodiment, the optical data modulator 130 is a single-input zero-sound Mach-Zehnder modulator (not shown). In this embodiment, the zero-sound MZI modulator includes a single input coupled to one of a differential output 124 and a complementary differential output 126 of the differential amplifier 114. The zero-sound MZI modulator can be driven by one of the differential signal g⑴ and the complementary differential signal -g⑴. Compared with an optical data modulator operating at the same bit rate, the bandwidth of the optical data modulator 130 is lower because the four-level data signal used to drive the modulator 130 has a lower frequency. Wide relationship. Therefore, the transmitter 100 of the present invention can use a relatively inexpensive optical data modulator, thereby reducing the overall cost of the optical data transmitter 100. In one of the embodiments, the design of the bias point of the optical data modulator 130 can eliminate the need to use a separate bias voltage. In another embodiment, the optical data modulator 130 includes a separate input (not shown) for applying a bias voltage to the modulator 130. In this embodiment, a separate bias voltage source (not shown) is applied to the input 21 200302934 in order to set the operating point of the modulator. In operation, the optical transmitter 100 receives a series of binary data bit (d) streams at a bit rate of 1 / T at the serial data input 104 of the dual binary pre-encoder 102, where T is This bit period. The binary binary pre-encoder 102 encodes the binary data bits (d). The dual binary pre-encoder 102 generates a binary pre-encoded bit sequence m and a complementary binary pre-encoded bit sequence 5 at the output 106 and the complementary output 108 of the binary binary pre-encoder 102, respectively. . In one embodiment, the binary precoding bit sequence m can be generated according to the following formula: mfmM ㊉ & (2) where i is a bit index and ㊉ is a logical mutex or (XOR) operand. The time-varying electrical signals m⑴ and ^ t) correspond to the binary precoding bit sequence m and the complementary binary precoding bit sequence 叾, respectively. The delay element 110 will delay the complementary pre-coded data signal with respect to the pre-coded data signal m in time. In one embodiment, the delay element 110 substantially generates a delay length between 0.4T and 0.9T in time relative to the precoded data signal m 叾 ⑴. The optimal time delay depends on the bandwidth of the signals and the linearity of the differential amplifier 114. For example, when the bandwidth is seventy-five percent (75%) of the bit rate and a linear differential amplifier, the optimal time delay is about 0.75T. The differential amplifier 114 receives the pre-encoded data signal m (t) and the delayed complementary pre-encoded data signal 5 (t), and sets the binary first 22 200302934 to the encoded data signal m⑴ and the complementary binary pre-encoded signal. The data signal 转换 is converted into a four-level data signal g⑴ above the differential output 124 and a complementary four-level data signal -g⑴ above the complementary differential output 126, respectively. The four-level data signal g⑴ is expressed as follows: g⑴ = k (m⑴-5 (t- τ)) (3)
其中k是該放大器的增益。於其中一實施例中，該差 動放大器114並非AC耦合。於此實施例中，該差動放大 器114包括一 DC偏壓位準，因此該四位準資料信號 便會變成g(t)+V offset 5 而該互補四位準資料信號則會變成- g⑴+ V〇ffSet 〇 於圖中所示的實施例中，該T型偏壓源128會增加一 偏壓電壓Vbias給該互補四位準資料信號-g⑴，用以產生一 含有DC偏移偏壓電壓之互補四位準資料信號-g(t)+Vbias。 因此，該差動放大器114與該T型偏壓源128便會產生兩 個信號，即該四位準資料信號g⑴以及該含有DC偏移偏壓 電壓之互補四位準資料信號-g(t)+Vbias。 該四位準資料信號g(t)以及該含有DC偏移偏壓電壓之 互補四位準資料信號-g(t)+Vbias與雙二進制編碼信號的差異 在於其總共有四個位準，也就是該等信號除了具有最大與 最小位準之外，還具有兩個中間位準。該等中間位準的形 狀就如同峰値出現於該等位元槽之邊界處的漣波。 該差動輸入（MZI)資料調變器130會接收雷射138所產 生的光學信號。該資料調變器130亦會從該差動放大器114 與該T型偏壓源128接收該四位準資料信號g(t)以及該含 23 200302934 有DC偏移偏壓電壓之互補四位準資料信號-g(t)+Vbias。該 資料調變器130會產生一經過調變的光學資料信號。 於其中一實施例中，該差動放大器Π4的增益經過選 擇之後，會使得該差動輸入MZI資料調變器130所接收到 的g(t)之峰至峰振幅變成Vtt。於另一實施例中，當該資料 調變器爲單輸入零響聲的調變器（圖中未顯示）時，該差動放 大器114的增益經過選擇之後則會使得g⑴之峰至峰振幅 變成2V;r。該些實施例對應的便係經由該調變器的最大光 學傳輸。 由該傳輸器1〇〇所傳輸的信號可藉由二進制強度直接 檢波法直接進行解碼或還原。將所接收到的光學振幅信號 平方之後便可達到二進制強度直接檢波的目的。因爲對光 學振幅進行平方係一平方律光學檢波器之固有功能，因此 並不需要於該接收器處提供特殊的處理功能便可還原該原 始的二進制資料序列d。 圖2A顯示的係可配合本發明之容忍色散之傳輸器使用 的MZI資料調變器之被傳輸光學振幅傳輸特徵200。該 MZI資料調變器可能是一單輸入零響聲的MZI資料調變器 或是一差動輸入MZI資料調變器。 根據本發明，該MZI資料調變器係由一四位準輸入資 料信號202加以驅動。該四位準輸入資料信號202可能是 直接被送至一單輸入零響聲的MZI調變器的輸入處，或是 藉由將信號送至該差動輸入MZI資料調變器之兩個輸入處 之後而產生出來的一差動信號。該MZI資料調變器會產生 24 200302934 一經過調變的光學振幅資料信號a(t) 204，其爲一雙極性信 號並且包括四個振幅位準。該調變器偏壓電壓經過選擇之 後，當該四位準輸入資料信號202等於包含該四位準輸入 資料信號202之該等四個位準之平均値時，該MZI資料調 變器的光學輸出便會成爲零値。 圖2B顯示的係可運用於本發明之容忍色散之傳輸器中 的MZI資料調變器之被傳輸光學強度傳輸特徵。該四位準 輸入資料信號202可能是直接被送至一單輸入零響聲的 MZI調變器的輸入處，或是藉由將信號送至該差動輸入 MZI資料調變器之兩個輸入處之後而產生出來的一差動信 號。該MZI資料調變器會產生一光學強度資料信號I(t) 212 ，其於零強度處具有最小値並且與該光學振幅資料信號a(t) 204(圖2A)的平方成正比。 圖3顯示的係圖1中容忍色散之傳輸器100作業的代 表性位元序列與信號300。該等位元序列與信號300係代表 一會產生0.75T之延遲時間的延遲元件110(圖1)，該0.75T 之延遲時間對應的係頻寬約爲該位元率之百分七十五（75%) 時的最佳時間延遲。 代表性二進制資料位元圖形302顯示的係位元率等於 1/T時（T爲位元週期），二進制資料位元（1„的代表性信號流 。二進制前置編碼資料圖形304顯示的係經過該雙二進制 前置編碼器1〇2(圖1)處理之後，對應於二進制資料位元圖 形302中之二進制資料位元序列d的二進制前置編碼位元 序列m。互補二進制前置編碼資料圖形306顯示的係經過 25 200302934 該雙二進制前置編碼器102處理之後，對應於二進制資料 位元圖形302中之二進制資料位元dn的互補二進制前置編 碼位元序列叾。 時間變動電氣信號圖形308顯示的係對應該二進制前 置編碼位元序列m的時間變動電氣信號m⑴。延遲互補時 間變動電氣信號圖形310表示的係經過延遲元件110(圖1) 延遲之後的互補時間變動電氣信號。 四位準資料信號圖形312顯示的係由該差動放大器 114(圖1)於該差動輸出124處所產生的四位準資料信號 g(t)。互補四位準資料信號圖形314顯示的係由該差動放大 器114於該互補差動輸出126處所產生的互補四位準資料 信號-g(t)。該資料信號g⑴以及該互補資料信號-g⑴都係四 位準資料信號，因爲經由該延遲元件110所產生的延遲都 小於一個完整的位元週期。如果經由該延遲元件11〇(圖1) 所產生的延遲等於一個完整的位元週期的話（圖中未顯示）， 那麼該資料信號g⑴以及該互補資料信號-g⑴便都係三位準 之雙二進制信號。 調變後光學振幅圖形316顯示的係響應該四位準資料 信號g⑴以及該互補四位準資料信號-g⑴而調變之後的調變 光學信號之光學振幅a(t)。該調變光學信號之光學振幅爲一 雙極性信號並且包括四個位準。 於其中一實施例中，該光學資料調變器130(圖1)包括 一經過選擇的預設操作點，當該四位準資料信號的振幅實 質等於該四位準資料信號之該等四個位準的平均値時，該 26 200302934 調變後的光學信號之強度會最小。於其中一實施例中，該 光學資料調變器130包括一經過選擇的預設操作點，當該 四位準資料信號的振幅實質等於該四位準資料信號之該等 四個位準的平均値，而且該互補四位準資料信號的振幅實 質等於該互補四位準資料信號之該等四個位準的平均値時 ，該調變後的光學信號之強度會最小。 該光學振幅信號之該等兩個中間位準的峰値係出現於 該等位元槽之邊界處。該些峰値的符號會不斷地交替，因 此當於色散的影響下而增寬時，其振幅平均値會等於零値 。峰値交替變換的優點在於可以保留相鄰1位準（即最大與 最小位準）的振幅，並且保留零位準的信號品質。如此便可 較先前技術之三位準雙二進制傳輸方式大幅地改善對色散 的容忍能力。 光學強度圖形318顯示的係響應該四位準資料信號以 及該互補四位準資料信號而調變之後的光學信號之光學功 率I(t)。該光學功率I⑴會與該光學振幅a⑴的平方成正比 。該光學功率I(t)爲一三位準信號。 由該傳輸器1〇〇所傳輸的信號可藉由二進制強度直接 檢波法直接進行解碼或還原。將所接收到的光學振幅信號 平方之後便可達到二進制強度直接檢波的目的。因爲對光 學振幅進行平方係一平方律光學檢波器之固有功能，因此 並不需要於該接收器處提供特殊的處理功能便可還原該原 始的二進制資料序列d。 此外，圖3中所示的代表性位元序列與信號包括一用 27 200302934 以闡述被接收信號S⑴的接收信號圖形320 ’其代表的係利 用150km長、具16ps/(km.nm)色散係數的光纖進行傳播之 後，於一無雜訊檢波器中所接收到之電氣輸出。與利用雙 二進制信號處理方式進行編碼的信號比較起來’該被接收 信號s⑴於色散容忍能力方面表現出明顯的改善程度。被接 收位元序列圖形322顯示的則係對應該被接收信號s⑴之被 接收位元序列。 圖4A顯示的係被接收信號的模擬l〇Gb/S光學眼圖 350，該信號係以圖1中具有一可產生一等於一個完整位元 週期之延遲（l〇〇ps延遲）之延遲元件11〇的傳輸器1〇〇，利 用150km長、16ps/(km.mn)色散的光纖進行傳輸。與習知 的先前技術之傳輸器比較起來，該傳輸器1〇〇具有改良的 色散容忍能力。不過，該光學眼圖350中的開口比較窄， 表示會有大量的色散損失。不良的色散容忍能力會表現於 符號間的干擾中，而且該些干擾並非因爲各種組件（例如光 纖放大器）所產生之光學雜訊累積之後所造成的，因爲此等 效應並未涵蓋於模擬圖中。將該延遲元件110所產生的延 遲時間降低至小於一個完整位元週期便可改良圖1之傳輸 器1〇〇的色散容忍能力。 圖4B顯示的係接收到信號的模擬lOGb/s光學眼圖 352，該信號係以圖1中具有可產生0.75T延遲（75ps延遲） 之延遲元件110的容忍色散之傳輸器100，利用150km長 、16ps/(km.nm)色散的光纖進行傳輸。與該使用完整位元週 期之延遲時間的傳輸器100比較起來，該使用小於一個位 28 200302934 元週期之延遲時間的容忍色散之傳輸器100會在光學眼圖 352中產生較大的開口。與圖4A中較窄的光學眼開口比較 起來，圖4B中較大的光學眼開口表示的係已經改良對色散 的容忍能力，如此便可降低接收器處的位元錯誤率。 因此，與使用雙二進制編碼處理之先前技術所產生的 三位準信號比較起來，該容忍色散之傳輸器1〇〇所產生的 四位準信號經過傳播之後可產生較大的光學眼。所以，與 習知的NRZ光纖傳輸器及先前技術中的雙二進制傳輸器比 較起來，該容忍色散之傳輸器1〇〇對於光纖色散效應及非 線性現象效應的容忍能力比較強。 如前面所討論般，色散容忍能力的改良其至少部分原 因係因爲該延遲元件11〇(圖1)所產生的延遲小於一個完整 的位元週期而使得該光學振幅會以零値爲中心進行震盪。 當該信號沿著該條光纖進行傳播時，該等震盪結果會趨於 衰減成零値，因而便可保留1位準的振幅，而且不會損及 零位準的信號品質。該等震盪的峰値係出現在該等位元的 邊界處，因此該等震盪並不會損及消失率。在先前技術具 有完整位元週期之延遲時間的雙二進制信號中，則不會出 現該些震盪情形。 爲能產生最高的色散容忍能力，該延遲元件11〇(圖1) 所產生的最佳延遲係取決於該信號頻寬。一般來說，當延 遲降低或信號頻寬增加時，該等震盪的峰値振幅便會提高 。舉例來說，當將信號頻寬限制在該位元率之百分之七十 五（75%)時，該延遲元件110所產生的最佳延遲時間便約爲 29 200302934 0.75T。如此便可使得該四位準信號之中間位準的振幅約爲 該四位準信號之對應峰値振幅的百分之五十（50%)。 圖5顯示的係根據本發明之容忍色散之傳輸器400的 一具體實施例，其包括一平行資料輸入406。該平行輸入之 容忍色散之傳輸器400與配合圖1所述之串列輸入之容忍 色散之傳輸器10〇相似。 不過，該傳輸器400包栝一平行雙二進制前置編碼器 402及一分時多工器404。該平行雙二進制前置編碼器402 _ 包括一平行輸入資料匯流排406，其寬度等於w ;而資料輸 入則爲V至d(w-1}。該平行編碼率則等於i/w乘以等效串列 資料流之位元率。 該平行雙二進制前置編碼器402會以平行的方式對資 料匯流排406上的資料進行編碼。與配合圖1所述之串列 輸入容忍色散之傳輸器100比較起來，平行編碼率的速率 較低。因爲可以場可程式化閘極陣列或特定應用積體電路 (ASIC)來實現該平行雙二進制前置編碼器4〇2，所以其成本 鲁 比較低廉。利用可程式化閘極陣列或ASIC可大幅地降低該 平行輸入傳輸器400的成本。 該平行雙二進制前置編碼器4〇2可產生一寬輸出408 ’ 其表示方式如下： j3[l，w] (4) cX;㊉ d „ 其中η爲低速平行資料流之取樣數，而w則逶輸入通 道之數量。 30 200302934 5亥平行雙一進制即置編碼器之w寬輸出4〇8會被親合 至分時多工器404。該分時多工器4〇4會分別在輸出41〇之 上產生一二進制前置編碼位元序列m以及在互補輸出412 之上產生一互補二進制前置編碼位元序列^，其中 mfcS:) (5) 其中int(i/w)係當位元率比被送至該平行雙二進制前置 編碼器402之資料輸入dQ至d(wA的資料信號之位元率快w 倍時，i/w之商數的整數部份。時間變動電氣信號m⑴及 ^(t)則分別對應該二進制前置編碼位元序列m及該互補二 進制前置編碼位元序列5，就如同配合圖1之傳輸器1〇〇所 述般。 該平行輸入傳輸器400包括一延遲元件11〇，其具有一 被耦合至該分時多工器404之互補輸出412的輸入112。此 外，該平行輸入傳輸器400包括一差動放大器114，其具有 一被耦合至該延遲元件110之輸出120的第一輸入118,以 及一被親合至該分時多工器404之輸出410的第二輸入122 〇 於另一實施例中（圖中未顯示），該延遲元件Π0係被耦 合於該分時多工器404之輸出410與該差動放大器114之 第二輸入122之間。於此實施例中，該分時多工器404之 互補輸出412則係被直接耦合至該差動放大器Π4之第一 輸入118。 於另一實施例中（圖中未顯示），該延遲元件11 〇係被耦 合於該分時多工器404之互補輸出412與該差動放大器114 31 200302934 之第一輸入118之間。第二延遲元件（圖中未顯示）則係被耦 合於該分時多工器404之輸出410與該差動放大器114之 第二輸入122之間。 該差動放大器114可將該二進制前置編碼資料信號以 及該互補二進制前置編碼資料信號分別轉換成差動輸出124 之上的差動信號g⑴以及互補差動輸出126之上的互補差 動信號-g⑴。該差動信號g(t)以及該互補差動信號-g⑴都係 四位準資料信號。 於其中一實施例中，存在一偏壓網路（例如T型偏壓源 128)，用以增加一偏壓電壓給該差動信號g⑴以及該互補差 動信號-g⑴中其中一者。於圖中所示的實施例中，該T型 偏壓源128係增加一偏壓電壓給該互補差動信號-g⑴，用 以產生一含有DC偏移電壓之互補差動信號-g(t)+Vbias。 此外，該平行輸入傳輸器400包括一光學資料調變器 130。於圖中所示的實施例中，該光學資料調變器130係一 差動輸入的馬赫-曾德爾干涉（MZI)資料調變器，其具有分 別被耦合至該差動放大器114之差動輸出124與互補差動 輸出126的第一資料輸入132與第二資料輸入134。 該光學資料調變器130亦包括一光學輸入136，用以從 一光源（例如雷射138)接收一光學信號。該資料調變器130 可響應該四位準資料信號以及該互補四位準資料信號對一 連續波光學信號的振幅進行調變，並且產生一經過調變的 光學輸出信號。 如配合圖1之傳輸器1〇〇所述般，由該平行輸入傳輸 32 200302934 器400所傳輸的信號可以利用二進制強度直接檢波法之接 收器（圖中未顯示）進行還原。會有一（l:w)分時解多工器（圖 中未顯示）對該等被檢測到之信號進行解多工處理，以還原 該輸入資料信號d。 _ 6顯示的係可供圖5之平行輸入傳輸器400使用的 四位元寬（w=4)平行雙二進制前置編碼器電路450之一具體 實施例的功能方塊圖。該平行雙二進制前置編碼器電路450 包括四個D型正反器452。每個該等正反器452都會於資 料輸入454處接收該等四個輸入資料信號…至d3中其中一 者。每個該等正反器452都具有一被連接至一共用時脈457 的時脈輸入456。 每個該等正反器452之互補輸出458會被連接至一互 斥或閘462之其中一個輸入460。每個該等互斥或閘462之 輸出464會被連接至四個D型正反器468的資料輸入466 。每個該等互斥或閘462之另一個輸入470則會被連接至 該等互斥或閘462中其中一個之輸出464或是被連接至該 等四個D型正反器468中其中一個之輸出472。 每個該等四個D型正反器468之輸出472都可能會被 耦合至一緩衝器474。該四位元寬（w=4)平行雙二進制前置 編碼器電路450具有一四位元寬之平行雙二進制資料輸出 476，用以輸出g亥等輸出資料信號c〇至c3。圖5之平行輸 入傳輸器400可以使用各種類型的平行雙二進制前置編碼 器電路。 圖7A顯示的係根據本發明之容忍色散之傳輸器5〇〇的 33 200302934 另一具體實施例，其包括一串列資料輸入。圖7A之傳輸器 500與圖1之傳輸器100類似。不過，於此實施例中，於該 等輸入及該差動放大器之間並沒有任何的淨延遲。 該傳輸器500包括一如圖1所述之串列雙二進制前置 編碼器102。該雙二進制前置編碼器102具有一串列資料輸 入104，其可接收一輸入資料信號。該雙二進制前置編碼器 102亦具有一輸出106以及一互補輸出108，用以分別產生 一二進制前置編碼資料信號以及一互補二進制前置編碼資 料信號。 該雙二進制前置編碼器102的輸出106係被連接至該 差動放大器114之第二輸入122。該雙二進制前置編碼器 102的互補輸出108係被連接至該差動放大器114之第一輸 入118。於其它實施例中，則會將一單輸入的放大器（圖中 未顯示）連接至該雙二進制前置編碼器102之輸出106及互 補輸出108中其中一者。於本文中配合圖7B所述之其它實 施例中，並未使用該差動放大器114，而該雙二進制前置編 碼器102則會產生輸出電壓足以直接驅動一調變器的信號 〇 該差動放大器114可將該二進制前置編碼資料信號以 及該互補二進制前置編碼資料信號分別轉換成差動輸出124 之上的差動信號g⑴以及互補差動輸出126之上的互補差 動信號_g(t)。 於其中一實施例中，存在一偏壓網路（例如T型偏壓源 128)，用以增加一偏壓電壓給該差動信號g⑴以及該互補差 34 200302934 動信號-g⑴中其中一者。於圖中所示的實施例中，該τ型 偏壓源128係增加一偏壓電壓給該互補差動信號-g⑴，用 以產生一含有DC偏移電壓之互補差動信號-g(t)+Vbias。於 另一實施例中，該差動放大器114則包括一偏壓電壓，用 以增加一偏壓電壓給該差動信號g⑴以及該互補差動信號-g(t)中其中一者。 該傳輸器500包括一延遲元件110，其具有一被耦合至 該差動放大器114之互補差動輸出12 6的輸入112。於另一 實施例中，該延遲元件110則係被耦合至該差動放大器114 之差動輸出124。於另一實施例中，則會有一第二延遲元件 (圖中未顯示）被耦合至該差動放大器114之差動輸出124。 該延遲元件110會將該含有DC偏移電壓之互補差動 信號-g(t)+Vbias相對於該差動信號g⑴延遲一段時間r，用 以產生一含有DC偏移電壓之延遲互補差動信號-g(t-r )+Vbias。於其中一實施例中，τ表示的係對應於小於該二 進制前置編碼資料信號之一個位元週期的時間。於其中一 實施例中，τ*的範圍介於0.4Τ至0.8Τ之間。爲能產生最 高的色散容忍能力，該延遲元件110所產生的最佳延遲係 與該等信號g⑴及-g⑴的頻寬成函數關係。舉例來說，當該 些信號的頻寬爲該位元率之百分之七十五（75%)時，最佳的 延遲時間便約爲0.6T。 此外，該傳輸器500包括一光學資料調變器130。於圖 中所示之實施例中，該光學資料調變器130係一差動輸入 MZI資料調變器。該差動輸入資料調變器130之第一資料 35 200302934 輸入132係被耦合至該差動放大器114之差動輸出124 ;該 差動輸入資料調變器130之第二資料輸入134則係被耦合 至該延遲元件110的輸出120。於其中一實施例中，當該延 遲元件110係被耦合至該差動放大器114之輸出124時， 該差動輸入資料調變器130之第一資料輸入132便會被耦 合至該延遲元件110的輸出120，而該差動輸入資料調變器 130之第二資料輸入134則會被耦合至該差動放大器114的 互補差動輸出126。 於其中一實施例中，則可將該延遲元件110整合至一 條纜線（圖中未顯示）中，該條纜線可將該差動放大器114之 互補輸出126耦合至該差動輸入資料調變器130之第二資 料輸入134。於另一實施例中，該延遲元件110則包括該條 纜線本身，透過該條纜線來傳播該互補二進制前置編碼資 料信號便可產生該延遲。於此實施例中，該條纜線的長度 必須經過選擇以對應該延遲。 該差動輸入資料調變器130亦包括一光學輸入136，用 以從一光源（例如雷射138)接收一光學信號。該資料調變器 130可響應該差動信號g⑴以及該含有DC偏移電壓之延遲 互補差動信號-g(t-r )+Vbias對一連續波光學信號進行調變 。於其中一實施例中，該偏壓電壓Vbias經過調整之後，當 g(t)+g(t-r)-Vbias 等於 g⑴+8〇1)-¥13&最大値及 g⑴+g(t-r )-Vbias最小値的平均値時，該差動輸入資料調變器130 所輸出的光學輸出功率便會處於最小的功率位準處。 圖7B顯示的係根據本發明之容忍色散之傳輸器550的 36 200302934 另一具體實施例，其包括一串列資料輸入。圖7B之傳輸器 550與圖7A之傳輸器500類似。不過，於此實施例中，該 傳輸器550並不包括一差動放大器114。該串列雙二進制前 置編碼器552則會產生一信號，用以直接驅動一光學資料 調變器，而不必使用任何的外部放大作用。 該串列雙二進制前置編碼器552具有一串列資料輸入 104，其可接收一輸入資料信號。該雙二進制前置編碼器 552亦具有一輸出554以及一互補輸出556，用以分別產生 一差動信號g⑴以及一互補差動信號-g⑴。於其中一實施例 中，該串列雙二進制前置編碼器552包括一內部放大器（圖 中未顯示），用以將該等差動信號g⑴以及互補差動信號-g(t)放大至足以直接驅動一光學調變器之位準。 於其中一實施例中，存在一偏壓網路（例如T型偏壓源 128)，用以增加一偏壓電壓給該差動信號g⑴以及該互補差 動信號-g⑴中其中一者。於圖中所示的實施例中，該T型 偏壓源128係增加一偏壓電壓給該互補差動信號-g⑴，用 以產生一含有DC偏移電壓之互補差動信號-g(t)+Vbias。 該傳輸器550包括一延遲元件110，其具有一被耦合至 該雙二進制前置編碼器552之互補輸出556的輸入112。該 延遲元件110會將該含有DC偏移電壓之互補差動信號-g(t)+Vbias相對於該差動信號g⑴延遲一段時間r，用以產 生一含有DC偏移電壓之延遲互補差動信號-g(t-r )+Vbias。 於其中一實施例中，r表示的係對應於小於該二進制前置 編碼資料信號之一個位元週期的時間。 37 200302934 此外，該傳輸器550包括一差動輸入MZI資料調變器 130。該差動輸入資料調變器130之第一資料輸入132係被 耦合至該雙二進制前置編碼器552之輸出554 ;該差動輸入 資料調變器130之第二資料輸入134則係被耦合至該延遲 元件110的輸出120。 該差動輸入資料調變器130亦包括一光學輸入136，用 以從一光源（例如雷射138)接收一光學信號。該資料調變器 130可響應該差動信號g⑴以及該含有DC偏移電壓之延遲 互補差動信號-g(t-r )+Vbias對一連續波光學信號進行調變 。於其中一實施例中，該偏壓電壓Vbias經過調整之後，當 g(t)+g(t-r)-Vbias 等於 g⑴+8〇〇-¥13&最大値及 g⑴+g(t-r )-Vbias最小値的平均値時，該差動輸入資料調變器130 所輸出的光學輸出功率便會處於最小的功率位準處。 圖8顯示的係圖7A中容忍色散之傳輸器500作業的代 表性位元序列與信號550。該等位元序列與信號550係代表 一會產生0.6T之延遲時間的延遲元件110，該0.6T之延遲 時間對應的係頻寬約爲該位元率之百分七十五（75%)時的最 佳時間延遲。該等用以顯示圖7A之容忍色散之傳輸器500 作業情形的代表性位元序列與信號550與該等用以顯示圖1 之容忍色散之傳輸器1〇〇作業情形的代表性位元序列與信 號300類似，不同的地方僅在於該等位元序列與信號550( 即g⑴與-g(t-r ))都係二進制信號，而非四位準的信號。 代表性二進制資料位元圖形552顯示的係位元率等於 1/T時（T爲位元週期），二進制資料位元1的代表性信號流 38 200302934 。二進制前置編碼資料圖形554顯示的係經過該雙二進制 前置編碼器1〇2(圖7A)處理之後，對應於二進制資料位元 圖形552中之二進制資料位元序列d的二進制前置編碼位 元序列m。互補二進制前置編碼資料圖形556顯示的係經 過該雙二進制前置編碼器102(圖7A)處理之後，對應於二 進制資料位元圖形552中之二進制資料位元dn的互補二進 制前置編碼位元序列G。Where k is the gain of the amplifier. In one embodiment, the differential amplifier 114 is not AC-coupled. In this embodiment, the differential amplifier 114 includes a DC bias level, so the four-level data signal becomes g (t) + V offset 5 and the complementary four-level data signal becomes-g⑴ + V〇ffSet 〇 In the embodiment shown in the figure, the T-shaped bias source 128 will add a bias voltage Vbias to the complementary four-level data signal -g⑴ to generate a bias voltage containing DC offset. The complementary four-level data signal of voltage -g (t) + Vbias. Therefore, the differential amplifier 114 and the T-type bias source 128 will generate two signals, namely the four-level data signal g⑴ and the complementary four-level data signal -g (t ) + Vbias. The four-level data signal g (t) and the complementary four-level data signal -g (t) + Vbias containing the DC offset bias voltage are different from the dual binary coded signal in that they have a total of four levels. That is, in addition to the maximum and minimum levels, these signals also have two intermediate levels. The shapes of these intermediate levels are like ripples that appear at the boundaries of these bit slots. The differential input (MZI) data modulator 130 receives an optical signal generated by the laser 138. The data modulator 130 will also receive the four-level data signal g (t) and the complementary four-level with a DC offset bias voltage of 23 200302934 from the differential amplifier 114 and the T-type bias source 128. Data signal -g (t) + Vbias. The data modulator 130 generates a modulated optical data signal. In one embodiment, after the gain of the differential amplifier Π4 is selected, the peak-to-peak amplitude of g (t) received by the differential input MZI data modulator 130 becomes Vtt. In another embodiment, when the data modulator is a single-input zero-sound modulator (not shown in the figure), the gain of the differential amplifier 114 is selected so that the peak-to-peak amplitude of g 变成 becomes 2V; r. These embodiments correspond to the maximum optical transmission through the modulator. The signal transmitted by the transmitter 100 can be directly decoded or restored by the binary intensity direct detection method. After the received optical amplitude signal is squared, the purpose of binary intensity direct detection can be achieved. Because of the inherent function of square-to-square-law optical detectors for squaring optical amplitudes, it is not necessary to provide special processing functions at the receiver to restore the original binary data sequence d. Figure 2A shows the transmitted optical amplitude transmission characteristic 200 of an MZI data modulator that can be used with the dispersion-tolerant transmitter of the present invention. The MZI data modulator may be a single-input zero-sound MZI data modulator or a differential input MZI data modulator. According to the present invention, the MZI data modulator is driven by a four-level input data signal 202. The four-level quasi-input data signal 202 may be sent directly to the input of a single-input zero-sound MZI modulator, or by sending the signal to two inputs of the differential-input MZI data modulator. A differential signal is generated afterwards. The MZI data modulator generates 24 200302934 a modulated optical amplitude data signal a (t) 204, which is a bipolar signal and includes four amplitude levels. After the modulator bias voltage is selected, when the four-level input data signal 202 is equal to the average of the four levels including the four-level input data signal 202, the optical of the MZI data modulator The output becomes zero. Figure 2B shows the transmitted optical intensity transmission characteristics of an MZI data modulator that can be used in the dispersion-tolerant transmitter of the present invention. The four-level quasi-input data signal 202 may be sent directly to the input of a single-input zero-sound MZI modulator, or by sending the signal to two inputs of the differential-input MZI data modulator. A differential signal is generated afterwards. The MZI data modulator generates an optical intensity data signal I (t) 212 which has a minimum chirp at zero intensity and is proportional to the square of the optical amplitude data signal a (t) 204 (FIG. 2A). FIG. 3 shows a representative bit sequence and signal 300 of the dispersion-tolerant transmitter 100 operation of FIG. These bit sequences and signals 300 represent a delay element 110 (Figure 1) that will produce a delay time of 0.75T. The delay time corresponding to the 0.75T delay time is about 75% of the bit rate. (75%) at the best time delay. When the bit rate of the representative binary data bit pattern 302 is equal to 1 / T (T is the bit period), the binary data bit (1 „represents the signal flow. After processing by the dual binary pre-encoder 102 (FIG. 1), a binary pre-encoded bit sequence m corresponding to the binary data bit sequence d in the binary data bit pattern 302. Complementary binary pre-encoded data The figure 306 shows the sequence of complementary binary precoding bits corresponding to the binary data bit dn in the binary data bit pattern 302 after the processing of the binary binary precoder 102 of 2003200334. Time-varying electrical signal pattern The system shown at 308 corresponds to the time-varying electrical signal m⑴ of the binary precoding bit sequence m. The delayed complementary time-varying electrical signal pattern 310 represents the complementary time-varying electrical signal after delay by the delay element 110 (Fig. 1). The level data signal graph 312 shows the four-level data generated by the differential amplifier 114 (FIG. 1) at the differential output 124. Signal g (t). The complementary four-level data signal pattern 314 shows the complementary four-level data signal -g (t) generated by the differential amplifier 114 at the complementary differential output 126. The data signal g⑴ and The complementary data signals -g⑴ are all four-level data signals, because the delays generated by the delay element 110 are less than a full bit period. If the delay generated by the delay element 110 (Figure 1) is equal to one For a complete bit period (not shown in the figure), then the data signal g⑴ and the complementary data signal -g⑴ are three-level bi-binary signals. After modulation, the optical amplitude graph 316 shows the response to the four bits Quasi-data signal g⑴ and the complementary four-level data signal -g⑴ and the optical amplitude a (t) of the modulated optical signal after modulation. The optical amplitude of the modulated optical signal is a bipolar signal and includes four levels In one embodiment, the optical data modulator 130 (FIG. 1) includes a selected preset operating point. When the amplitude of the four-level data signal is substantially equal to that of the four-level data signal When the average level of the four levels is equal, the intensity of the 26 200302934 modulated optical signal will be the smallest. In one embodiment, the optical data modulator 130 includes a selected preset operating point. The amplitude of the four-level data signal is substantially equal to the average of the four levels of the four-level data signal, and the amplitude of the complementary four-level data signal is substantially equal to the four of the complementary four-level data signal. When the level is averaged, the intensity of the modulated optical signal will be the smallest. The peaks of the two intermediate levels of the optical amplitude signal appear at the boundaries of the bit slots. The peaks The sign of is constantly changing, so when widening under the influence of dispersion, its average amplitude 値 will be equal to zero 値. The advantage of the peak-to-peak alternate transformation is that it can preserve the amplitude of the adjacent 1 level (that is, the maximum and minimum levels) and the signal quality of the zero level. In this way, the tolerance to dispersion can be greatly improved compared with the three-bit quasi-binary binary transmission method of the prior art. The optical intensity graph 318 shows the optical power I (t) of the optical signal after being modulated in response to the four-level data signal and the complementary four-level data signal. The optical power I⑴ will be proportional to the square of the optical amplitude a⑴. The optical power I (t) is a three-level quasi-signal. The signal transmitted by the transmitter 100 can be directly decoded or restored by the binary intensity direct detection method. After the received optical amplitude signal is squared, the purpose of binary intensity direct detection can be achieved. Because of the inherent function of square-to-square-law optical detectors for squaring optical amplitudes, it is not necessary to provide special processing functions at the receiver to restore the original binary data sequence d. In addition, the representative bit sequence and signal shown in FIG. 3 includes a received signal pattern of 27 200302934 to illustrate the received signal S⑴ 320 '. Its representative system uses a 150km long, 16ps / (km.nm) dispersion coefficient After the fiber is propagated, the electrical output is received in a noise-free detector. Compared with signals encoded using a dual binary signal processing method, the received signal s shows a significant improvement in terms of dispersion tolerance. The received bit sequence pattern 322 shows the received bit sequence corresponding to the received signal s. FIG. 4A shows an analog 10 Gb / S optical eye diagram 350 of a received signal with a delay element in FIG. 1 having a delay (100 ps delay) equal to one full bit period. The 110 transmitter 100 uses a 150km long, 16ps / (km.mn) dispersion fiber for transmission. Compared to the conventional prior art transmitter, the transmitter 100 has improved dispersion tolerance. However, the opening in the optical eye diagram 350 is relatively narrow, indicating that there will be a large amount of dispersion loss. Poor dispersion tolerance is manifested in inter-symbol interference, and these interferences are not caused by the accumulation of optical noise generated by various components (such as fiber amplifiers), because these effects are not covered in the simulation diagram . Reducing the delay time generated by the delay element 110 to less than one full bit period can improve the dispersion tolerance of the transmitter 100 of FIG. 1. FIG. 4B shows an analog 10Gb / s optical eye diagram 352 of a received signal. The signal is a dispersion-tolerant transmitter 100 with a delay element 110 that can generate a 0.75T delay (75ps delay) in FIG. , 16ps / (km.nm) dispersion fiber for transmission. Compared with the transmitter 100 using the delay time of the full bit period, the dispersion-tolerant transmitter 100 using the delay time of less than one bit 28 200302934 yuan period will produce a larger opening in the optical eye diagram 352. Compared to the narrower optical eye opening in Fig. 4A, the larger optical eye opening in Fig. 4B indicates that the system has improved tolerance to dispersion, thus reducing the bit error rate at the receiver. Therefore, compared with the three-level signal generated by the prior art using bi-binary coding processing, the four-level signal generated by the dispersion-tolerant transmitter 100 can generate a larger optical eye after propagation. Therefore, compared with the conventional NRZ optical fiber transmitter and the dual binary transmitter in the prior art, the dispersion-tolerant transmitter 100 has a stronger tolerance to fiber dispersion effects and non-linear phenomenon effects. As discussed earlier, the improvement in dispersion tolerance is at least partly due to the delay produced by the delay element 11 (Fig. 1) being less than a full bit period, making the optical amplitude oscillate around zero chirp. . When the signal propagates along the fiber, these oscillation results tend to decay to zero chirp, so that the 1-level amplitude can be retained without compromising the signal quality of the zero-level signal. The peaks of these oscillations appear at the boundaries of these bits, so the oscillations will not damage the disappearance rate. In prior art dual binary signals with a delay time of a full bit period, these oscillations would not occur. In order to produce the highest dispersion tolerance, the optimal delay produced by the delay element 11 (Fig. 1) depends on the signal bandwidth. Generally, when the delay is reduced or the signal bandwidth is increased, the peak-to-peak amplitude of such oscillations increases. For example, when the signal bandwidth is limited to 75% (75%) of the bit rate, the optimal delay time generated by the delay element 110 is approximately 29 200302934 0.75T. In this way, the amplitude of the middle level of the four-level signal can be about fifty percent (50%) of the corresponding peak-to-peak amplitude of the four-level signal. FIG. 5 shows a specific embodiment of a dispersion-tolerant transmitter 400 according to the present invention, which includes a parallel data input 406. The parallel input dispersion-tolerant transmitter 400 is similar to the serial input dispersion-tolerant transmitter 100 described with reference to FIG. However, the transmitter 400 includes a parallel dual binary precoder 402 and a time division multiplexer 404. The parallel bi-binary pre-encoder 402_ includes a parallel input data bus 406, whose width is equal to w; and the data input is V to d (w-1}. The parallel encoding rate is equal to i / w multiplied by, etc. Effective bit rate of the serial data stream. The parallel bi-binary pre-encoder 402 encodes the data on the data bus 406 in parallel. In conjunction with the serial input dispersion-tolerant transmitter described in Figure 1 Compared with 100, the rate of parallel coding rate is lower. Because the field-programmable gate array or application-specific integrated circuit (ASIC) can be used to implement the parallel bi-binary precoder 40, its cost is relatively low. .Using a programmable gate array or ASIC can greatly reduce the cost of the parallel input transmitter 400. The parallel bi-binary pre-coder 40 can produce a wide output 408 'which is expressed as follows: j3 [l, w] (4) cX; ㊉ d „where η is the number of samples of the low-speed parallel data stream, and w is the number of input channels. 30 200302934 5 W-wide output of a parallel double-ary encoder Will be affixed to the time division multiplexer 404. The time division multiplexer 400 generates a binary precoding bit sequence m above the output 41 and a complementary binary precoding bit sequence ^ on the complementary output 412, where mfcS :) ( 5) where int (i / w) is i / w when the bit rate is w times faster than the data input dQ to d (wA of the data signal sent to the parallel bi-binary pre-encoder 402) The integer part of the quotient. The time-varying electrical signals m⑴ and ^ (t) correspond to the binary precoding bit sequence m and the complementary binary precoding bit sequence 5, respectively, just like the transmitter in conjunction with Figure 1. The parallel input transmitter 400 includes a delay element 110, which has an input 112 coupled to a complementary output 412 of the time division multiplexer 404. In addition, the parallel input transmitter 400 includes A differential amplifier 114 having a first input 118 coupled to the output 120 of the delay element 110 and a second input 122 coupled to the output 410 of the time division multiplexer 404. In the embodiment (not shown), the delay element Π0 is coupled to the branch Between the output 410 of the time multiplexer 404 and the second input 122 of the differential amplifier 114. In this embodiment, the complementary output 412 of the time division multiplexer 404 is directly coupled to the differential amplifier Π4 The first input 118. In another embodiment (not shown in the figure), the delay element 110 is coupled to the complementary output 412 of the time division multiplexer 404 and the differential amplifier 114 31 200302934. Between the inputs 118. A second delay element (not shown) is coupled between the output 410 of the time division multiplexer 404 and the second input 122 of the differential amplifier 114. The differential amplifier 114 can convert the binary pre-encoded data signal and the complementary binary pre-encoded data signal into a differential signal g⑴ above the differential output 124 and a complementary differential signal above the complementary differential output 126, respectively. -g⑴. The differential signal g (t) and the complementary differential signal -g⑴ are both four-level data signals. In one embodiment, there is a bias network (such as a T-shaped bias source 128) for adding a bias voltage to one of the differential signal g⑴ and the complementary differential signal -g⑴. In the embodiment shown in the figure, the T-shaped bias source 128 adds a bias voltage to the complementary differential signal -g⑴ to generate a complementary differential signal -g (t ) + Vbias. In addition, the parallel input transmitter 400 includes an optical data modulator 130. In the embodiment shown in the figure, the optical data modulator 130 is a Mach-Zehnder interference (MZI) data modulator with a differential input, which has differentials respectively coupled to the differential amplifier 114. The first data input 132 and the second data input 134 of the output 124 and the complementary differential output 126. The optical data modulator 130 also includes an optical input 136 for receiving an optical signal from a light source (e.g., a laser 138). The data modulator 130 can modulate the amplitude of a continuous wave optical signal in response to the four-level data signal and the complementary four-level data signal, and generate a modulated optical output signal. As described in conjunction with the transmitter 100 of FIG. 1, the signal transmitted by the parallel input transmission 32 200302934 can be restored using a binary intensity direct detection receiver (not shown). There will be a (l: w) time-division demultiplexer (not shown in the figure) to demultiplex the detected signals to restore the input data signal d. _ 6 is a functional block diagram of a specific embodiment of a four-bit wide (w = 4) parallel dual binary pre-encoder circuit 450 that can be used by the parallel input transmitter 400 of FIG. 5. The parallel dual binary pre-encoder circuit 450 includes four D-type flip-flops 452. Each of these flip-flops 452 will receive one of the four input data signals at data input 454 ... to d3. Each of these flip-flops 452 has a clock input 456 connected to a common clock 457. The complementary output 458 of each of these flip-flops 452 is connected to one of the inputs 460 of a mutex OR gate 462. The output 464 of each such mutex OR gate 462 is connected to the data input 466 of the four D-type flip-flops 468. The other input 470 of each of these mutexes or gates 462 is connected to the output 464 of one of the mutexes or gates 462 or to one of the four D-type flip-flops 468. The output is 472. The output 472 of each of these four D-type flip-flops 468 may be coupled to a buffer 474. The four-bit-wide (w = 4) parallel bi-binary pre-encoder circuit 450 has a four-bit-wide parallel bi-binary data output 476 for outputting output data signals such as g0 to c3. The parallel input transmitter 400 of Fig. 5 can use various types of parallel dual binary pre-encoder circuits. FIG. 7A shows another specific embodiment of a dispersion-tolerant transmitter 500 according to the present invention, which includes a series of data inputs. The transmitter 500 of FIG. 7A is similar to the transmitter 100 of FIG. However, in this embodiment, there is no net delay between the inputs and the differential amplifier. The transmitter 500 includes a tandem double-binary pre-encoder 102 as described in FIG. The bi-binary pre-encoder 102 has a serial data input 104 which can receive an input data signal. The dual binary precoder 102 also has an output 106 and a complementary output 108 for generating a binary precoded data signal and a complementary binary precoded data signal, respectively. An output 106 of the dual binary pre-encoder 102 is connected to a second input 122 of the differential amplifier 114. The complementary output 108 of the dual binary pre-encoder 102 is connected to the first input 118 of the differential amplifier 114. In other embodiments, a single-input amplifier (not shown) is connected to one of the output 106 and the complementary output 108 of the dual binary pre-encoder 102. In other embodiments described herein in conjunction with FIG. 7B, the differential amplifier 114 is not used, and the dual binary pre-encoder 102 will generate a signal with an output voltage sufficient to directly drive a modulator. The differential The amplifier 114 may convert the binary pre-encoded data signal and the complementary binary pre-encoded data signal into a differential signal g⑴ above the differential output 124 and a complementary differential signal _g ( t). In one embodiment, there is a bias network (such as a T-type bias source 128) for adding a bias voltage to the differential signal g⑴ and the complementary difference 34 200302934 dynamic signal -g⑴ . In the embodiment shown in the figure, the τ-type bias source 128 adds a bias voltage to the complementary differential signal -g⑴ to generate a complementary differential signal -g (t ) + Vbias. In another embodiment, the differential amplifier 114 includes a bias voltage for adding a bias voltage to one of the differential signal g⑴ and the complementary differential signal -g (t). The transmitter 500 includes a delay element 110 having an input 112 coupled to a complementary differential output 12 6 of the differential amplifier 114. In another embodiment, the delay element 110 is coupled to a differential output 124 of the differential amplifier 114. In another embodiment, a second delay element (not shown) is coupled to the differential output 124 of the differential amplifier 114. The delay element 110 delays the complementary differential signal -g (t) + Vbias containing the DC offset voltage relative to the differential signal g⑴ for a period of time r to generate a delayed complementary differential containing the DC offset voltage. Signal -g (tr) + Vbias. In one embodiment, τ represents a time corresponding to less than one bit period of the binary pre-encoded data signal. In one embodiment, τ * ranges from 0.4T to 0.8T. In order to produce the highest dispersion tolerance, the optimal delay generated by the delay element 110 is a function of the bandwidth of the signals g⑴ and -g⑴. For example, when the bandwidth of these signals is 75% (75%) of the bit rate, the optimal delay time is about 0.6T. In addition, the transmitter 500 includes an optical data modulator 130. In the embodiment shown in the figure, the optical data modulator 130 is a differential input MZI data modulator. The first data 35 200302934 input 132 of the differential input data modulator 130 is coupled to the differential output 124 of the differential amplifier 114; the second data input 134 of the differential input data modulator 130 is An output 120 is coupled to the delay element 110. In one embodiment, when the delay element 110 is coupled to the output 124 of the differential amplifier 114, the first data input 132 of the differential input data modulator 130 is coupled to the delay element 110. The second data input 134 of the differential input data modulator 130 is coupled to the complementary differential output 126 of the differential amplifier 114. In one embodiment, the delay element 110 can be integrated into a cable (not shown), and the cable can couple the complementary output 126 of the differential amplifier 114 to the differential input data modulation. The second data input 134 of the transformer 130. In another embodiment, the delay element 110 includes the cable itself, and the delay can be generated by propagating the complementary binary precoding data signal through the cable. In this embodiment, the length of the cable must be selected to correspond to the delay. The differential input data modulator 130 also includes an optical input 136 for receiving an optical signal from a light source (e.g., a laser 138). The data modulator 130 can modulate a continuous wave optical signal in response to the differential signal g⑴ and the delayed complementary differential signal -g (t-r) + Vbias containing a DC offset voltage. In one embodiment, after the bias voltage Vbias is adjusted, when g (t) + g (tr) -Vbias is equal to g⑴ + 8〇1)-¥ 13 & maximum 値 and g⑴ + g (tr) -Vbias When the minimum value is the average value, the optical output power output by the differential input data modulator 130 will be at the minimum power level. FIG. 7B shows another specific embodiment of the dispersion-tolerant transmitter 550 according to the present invention, which includes a series of data inputs. The transmitter 550 of Fig. 7B is similar to the transmitter 500 of Fig. 7A. However, in this embodiment, the transmitter 550 does not include a differential amplifier 114. The tandem binary binary pre-encoder 552 generates a signal for directly driving an optical data modulator without using any external amplification. The serial bi-binary pre-encoder 552 has a serial data input 104 which can receive an input data signal. The dual binary pre-encoder 552 also has an output 554 and a complementary output 556 for generating a differential signal g⑴ and a complementary differential signal -g⑴, respectively. In one embodiment, the tandem double-binary pre-encoder 552 includes an internal amplifier (not shown) for amplifying the differential signals g⑴ and the complementary differential signal -g (t) sufficiently. Directly drive the level of an optical modulator. In one embodiment, there is a bias network (such as a T-shaped bias source 128) for adding a bias voltage to one of the differential signal g⑴ and the complementary differential signal -g⑴. In the embodiment shown in the figure, the T-shaped bias source 128 adds a bias voltage to the complementary differential signal -g⑴ to generate a complementary differential signal -g (t ) + Vbias. The transmitter 550 includes a delay element 110 having an input 112 coupled to a complementary output 556 of the dual binary pre-encoder 552. The delay element 110 delays the complementary differential signal -g (t) + Vbias containing the DC offset voltage relative to the differential signal g⑴ for a period of time r to generate a delayed complementary differential containing the DC offset voltage. Signal -g (tr) + Vbias. In one embodiment, r is represented by a time corresponding to less than one bit period of the binary pre-encoded data signal. 37 200302934 In addition, the transmitter 550 includes a differential input MZI data modulator 130. The first data input 132 of the differential input data modulator 130 is coupled to the output 554 of the dual binary pre-encoder 552; the second data input 134 of the differential input data modulator 130 is coupled. To the output 120 of the delay element 110. The differential input data modulator 130 also includes an optical input 136 for receiving an optical signal from a light source (e.g., a laser 138). The data modulator 130 can modulate a continuous wave optical signal in response to the differential signal g⑴ and the delayed complementary differential signal -g (t-r) + Vbias containing a DC offset voltage. In one embodiment, after the bias voltage Vbias is adjusted, when g (t) + g (tr) -Vbias is equal to g⑴ + 800- ¥ 13 & maximum 値 and g⑴ + g (tr) -Vbias is minimum When the average of 値 is, the optical output power output by the differential input data modulator 130 will be at the minimum power level. Figure 8 shows a representative bit sequence and signal 550 of the dispersion tolerant transmitter 500 operation of Figure 7A. These bit sequences and signals 550 represent a delay element 110 that will produce a delay time of 0.6T. The delay time corresponding to the 0.6T delay time is about 75% (75%) of the bit rate. The best time delay. The representative bit sequences and signals 550 used to display the operation of the dispersion-tolerant transmitter 500 of FIG. 7A and the representative bit sequences used to display the 100 operation of the dispersion-tolerant transmitter 500 of FIG. 1 Similar to the signal 300, except that the bit sequence and the signal 550 (that is, g⑴ and -g (tr)) are binary signals, not four-level signals. When the bit rate of the representative binary data bit pattern 552 is equal to 1 / T (T is the bit period), the representative signal flow of the binary data bit 1 38 200302934. The binary pre-encoded data pattern 554 shows the binary pre-encoded bits corresponding to the binary data bit sequence d in the binary data bit pattern 552 after processing by the dual binary pre-encoder 102 (FIG. 7A). Metasequence m. The complementary binary pre-encoded data pattern 556 shows the complementary binary pre-encoded bit corresponding to the binary data bit dn in the binary data bit pattern 552 after being processed by the dual binary pre-encoder 102 (FIG. 7A). Sequence G.
差動信號圖形558顯示的係該差動放大器114(圖7A) 於差動輸出124處所產生的差動信號g(t)。延遲互補差動 信號圖形560顯示的係該差動放大器114以及可產生0.6T 延遲之延遲元件110所產生的延遲的互補差動信號-g(t-r) 〇 光學強度圖形562顯示的係響應該差動信號g⑴以及 該延遲互補差動信號-g(t- r )而調變之後的光學信號之光學 功率I(t)。與配合圖3之光學強度圖形318中所闡述的傳輸 器1〇〇所產生之光學功率波形I⑴相同，光學強度圖形562 所顯示的光學功率波形I⑴亦爲一三位準信號，其具有以該 等位元邊界爲中心的較小峰値。 此外，圖8中所示的代表性位元序列與信號包括一用 以闇述被接收信號s(t)的接收信號圖形564，其代表的係利 用150km長、具16ps/(km.nm)色散係數的光纖進行傳播之 後，於一無雜訊檢波器中所接收到之模擬電氣輸出。與利 用具有T延遲之先前技術中的方法所取得邊界信號處理方 式進行編碼的信號比較起來，圖8中該被接收信號s⑴於色 39 200302934 散容忍能力方面表現出明顯的改善程度。 圖9A顯示的係被接收到信號之模擬10Gb/s光學眼圖 600，該信號係以圖7A中具有一可產生一等於一個完整位 元週期之延遲（l〇〇ps延遲）之延遲元件110的傳輸器500， 利用150km長、16ps/(km.nm)色散的光纖進行傳輸。與先 前技術中習知的NRZ傳輸器比較起來，該傳輸器500具有 改良的色散容忍能力。不過，該光學眼圖600中呈現出較 爲閉合的形狀，表示利用150km長、16ps/(km.nm)色散的 光纖進行光學信號傳播之後，對色散的容忍較差。不良的 色散容忍能力會表現於符號間的干擾中，而且該些干擾並 非因爲各種組件（例如光纖放大器）所產生之光學雜訊累積之 後所造成的，因爲此等效應並未涵蓋於模擬圖中。將該延 遲元件110所產生的延遲時間降低至小於一個完整位元週 期便可改良圖7A之傳輸器500的色散容忍能力。如此作法 便可於圖8所示之光學功率波形562中的位元邊界處產生 具最大値的峰値。該些峰値的優點係可以藉由破壞於光纖 色散的影響下使得該等1位準增寬時而發生的干擾情形， 以保留連續1位準的振幅。 圖9B顯示的係被接收到信號的模擬lOGb/s光學眼圖 602，該信號係以圖7A中具有可產生0.6T延遲（60ps延遲） 之延遲元件110的容忍色散之傳輸器500，利用150km長 、16ps/(km.nm)色散的光纖進行傳輸。與該使用圖7A中完 整位元週期之延遲時間的傳輸器500比較起來，該使用小 於一個位元週期之延遲的容忍色散之傳輸器500會於該光 200302934 學眼關係圖602中產生較大的開口。該較大的光學眼開口 表示的係已經改良對色散的容忍能力，如此便可降低接收 器處的位元錯誤率。 圖1中傳輸器1〇〇的色散容忍能力大於圖7A中傳輸器 500的色散容忍能力。圖7A中傳輸器500的色散容忍能力 比較低的原因係因爲並非使用等振幅且相反符號之信號來 驅動該調變器130的關係。如此便會於該光學信號中造成 部份頻率聲響，而對傳播造成不利的影響。 圖10顯示的係根據本發明之容忍色散之傳輸器650的 另一具體實施例，其包括一平行資料輸入。該平行輸入容 忍色散之傳輸器650與配合圖7A所述之串列輸入容忍色散 之傳輸器500相似。 不過如配合圖5所述般，該傳輸器650包括一平行雙 二進制前置編碼器402及一分時多工器404。該平行雙二進 制前置編碼器402包括一平行輸入資料匯流排406，其寬度 等於w ;而資料輸入則爲f至。該平行雙二進制前置 編碼器402會以平行的方式對資料匯流排406上的資料進 行編碼。該平行編碼率則等於1/w乘以等效串列資料流之 位元率。 該平行雙二進制前置編碼器之w寬輸出408會被耦合 至分時多工器404。該分時多工器404會以串列資料的方式 分別在輸出410之上產生一二進制前置編碼位元序列m以 及在互補輸出412之上產生一互補二進制前置編碼位元序 列5，其位元率則比被送至該平行雙二進制前置編碼器 200302934 402之資料輸入V至d(w〜的資料信號之位元率快w倍。 該分時多工器404之輸出410與互補輸出412都係被 連接至一差動放大器114。於另一實施例中則會將一單輸入 的放大器（圖中未顯示）連接至該分時多工器404之輸出410 與互補輸出412。於其它實施例中，則未使用該差動放大器 114，而該分時多工器4〇4則會產生輸出電壓足以直接驅動 一調變器的信號。 該差動放大器Π4可將該二進制前置編碼資料信號以 及該互補二進制前置編碼資料信號分別轉換成差動輸出124 之上的差動信號g⑴以及互補差動輸出126之上的互補差 動信號-g⑴。該等差動信號g⑴以及互補差動信號-g⑴都係 頻寬有限的二位準資料信號。 於其中一實施例中，存在一偏壓網路（例如T型偏壓源 128)，用以增加一偏壓電壓給該差動信號g⑴以及該互補差 動信號-g⑴中其中一者。於圖中所示的實施例中，該T型 偏壓源128係增加一偏壓電壓給該互補差動信號_g⑴，用 以產生一含有DC偏移電壓之互補差動信號-g(t)+Vbias。於 另一實施例中’該差動放大器114則包括一偏壓電壓源， 用以增加一偏壓電壓給該差動信號g⑴以及該互補差動信 號-g⑴中其中一者。 該傳輸器650包括一延遲元件11〇，其具有一被賴合至 該差動放大器114之互補差動輸出126的輸入112。於另— 實施例中，該延遲元件110則係被耦合至該差動放大器114 之差動輸出124。於另一實施例中，則會有一第二延遲元件 42 200302934 (圖中未顯示）被耦合至該差動放大器Π4之差動輸出124。 該延遲元件110會將該含有DC偏移電壓之互補差動 信號-g(t)+Vbias相對於該差動信號g⑴延遲一段時間r，用 以產生一含有DC偏移電壓之延遲互補差動信號-g(t-r )+Vbias。於其中一實施例中，r表示的係對應於小於該二 進制前置編碼資料信號之一個位元週期的時間。於其中一 實施例中，r的範圍介於0.4T至0.8T之間。如配合圖5 所述般，爲能產生最高的色散容忍能力，該延遲元件110 所產生的最佳延遲係與該等信號g⑴及-g⑴的頻寬成函數關 係。舉例來說，當該些信號的頻寬爲該位元率之百分之七 十五（75%)時，最佳的延遲時間便約爲0.6T。 此外，該傳輸器650包括一光學資料調變器130。於圖 中所示之實施例中，該光學資料調變器130係一差動輸入 MZI資料調變器。該差動輸入資料調變器130之第一資料 輸入132係被耦合至該差動放大器114之差動輸出124 ;該 差動輸入資料調變器130之第二資料輸入134則係被耦合 至該延遲元件110的輸出120。 於另一實施例中，該延遲元件110則係被耦合至該差 動放大器Π4之輸出124。於此實施例中，該差動輸入資料 調變器130之第一資料輸入132會被耦合至該延遲元件110 的輸出120，而該差動輸入資料調變器130之第二資料輸入 134則會被耦合至該差動放大器114的互補差動輸出126。 該差動輸入資料調變器130亦包括一光學輸入136，用 以從一光源（例如雷射138)接收一光學信號。該資料調變器 43 200302934 130可響應該差動信號g⑴以及該含有DC偏移電壓之延遲 互補差動信號-g(t- r )+Vbias對一連續波光學信號進行調變 。於其中一實施例中，該偏壓電壓Vbias經過調整之後，當 g(t)+g(t- r )-Vbias 等於 g(t)+g(t- r )-Vbias 最大値及 g⑴+g(t-τ )-Vbias最小値的平均値時，該差動輸入資料調變器no 所輸出的光學輸出功率便會處於最小的功率位準處。 雖然已經詳細地參考較佳的實施例而特別顯示本發明 且予以說明，不過對於熟習本技術的人士而言應該瞭解的 係，在不脫離本文所界定之本發明精神與範疇下，可對其 進行各種形式上及細節上的修改。 【圖式簡單說明】 (一）圖式部分 配合圖式，便可非常淸楚本發明的優點，其中於不同 圖式中，相同的元件符號代表相同的結構元素與特徵。該 等圖式並不需要等比例縮放，反倒是在闡述本發明之原理 時還會加以放大。 圖1顯示的係根據本發明之容忍色散之傳輸器的一具 體實施例，其包括一串列資料輸入。 圖2A顯示的係可運用於本發明之容忍色散之傳輸器中 的MZI資料調變器之被傳輸光學振幅轉換特徵。 圖2B顯示的係可運用於本發明之容忍色散之傳輸器中 的MZI資料調變器之被傳輸光學強度轉換特徵。 圖3顯示的係圖1中容忍色散之傳輸器作業的代表性 200302934 位元序列與信號。 圖4A顯示的係爲被接收信號的模擬lOGb/s光學眼圖 ，該訊號係以圖1中具有一可產生一等於一個完整位元週 期之延遲（l〇〇ps延遲）之延遲元件的傳輸器，利用150km長 、16ps/(km.nm)色散的光纖進行傳輸。 圖4B顯示的係爲被接收信號的模擬lOGb/s光學眼圖 ，該信號係以圖1中具有可產生0.75T延遲（75ps延遲）之 延遲元件的容忍色散之傳輸器，利用150km長、 16ps/(km.nm)色散的光纖進行傳輸。 圖5顯示的係根據本發明之容忍色散之傳輸器的一具 體實施例，其包括一平行資料輸入。 圖6顯示的係四位元寬（w=4)平行雙二進制前置編碼器 電路之一具體實施例的功能方塊圖，其可配合圖5之平行 輸入之容忍色散之傳輸器使用。 圖7A顯示的係根據本發明之容忍色散之傳輸器的另一 具體實施例，其包括一串列資料輸入。 圖7B顯示的係根據本發明之容忍色散之傳輸器的另一 具體實施例，其包括一串列資料輸入。 圖8顯示的係圖7A中容忍色散之傳輸器作業的代表性 位元序列與信號。 圖9A顯示的係被接收信號的模擬lOGb/s光學眼圖， 該信號係以圖8中具有一可產生一等於一個完整位元週期 之延遲（l〇〇ps延遲）之延遲元件的傳輸器，利用150km長、 16ps/(km.nm)色散的光纖進行傳輸。 45 200302934 圖9B顯示的係被接收信號的模擬lOGb/s光學眼圖， 該信號係以圖8中具有可產生0.6T延遲（60ps延遲）之延遲 元件的傳輸器，利用150km長、16ps/(km.nm)色散的光纖 進行傳輸。 圖10顯示的係根據本發明之容忍色散之傳輸器的另一 具體實施例， 其包括一平行資料輸入。 (二）元件代表符號 100,500,550 包括串列資料輸入之容忍色散之傳輸器 102,552 串列雙二進制前置編碼器 104 串列雙二進制前置編碼器1〇2,552的串列 資料輸入 106 串列雙二進制前置編碼器102的輸出 108 串列雙二進制前置編碼器102的互補輸出 110 延遲元件 112 延遲元件110的輸入 114 差動放大器 118 差動放大器114的第一輸入 120 延遲元件Π0之輸出 122 差動放大器114的第二輸入 124 差動放大器Π4的差動輸出 126 差動放大器114的互補差動輸出 128 T型偏壓源 130 光學資料調變器 132 光學資料調變器130的第一資料輸入 200302934 134 135 136 137 138 400,650 402 404 406 408 410 412 450 452,468 454 456 457 458 460 462 464 466 470 光學資料調變器130的第二資料輸入 濾波器（第一濾波器） 資料調變器130的光學輸入 第二濾波器 雷射 包括平行資料輸入之容忍色散之傳輸器 平行雙二進制前置編碼器 分時多工器 容忍色散之傳輸器400的平行資料輸入 平行雙二進制前置編碼器402的輸出 分時多工器404之輸出 分時多工器404之互補輸出 四位元寬平行雙二進制前置編碼器電路 D型正反器 正反器452的資料輸入 正反器452的時脈輸入 共用時脈 正反器452的互補輸出 互斥或閘462之其中一個輸入 互斥或閘 互斥或閘462的輸出 正反器468的資料輸入 互斥或閘462的另一個輸入 正反器468的輸出 47 472 200302934 474 476 554 556 緩衝器 平行雙二進制前置編碼器電路450之資料 輸出 串列雙二進制前置編碼器552的輸出 串列雙二進制前置編碼器552的互補輸出The differential signal graph 558 shows the differential signal g (t) generated by the differential amplifier 114 (FIG. 7A) at the differential output 124. The delayed complementary differential signal graph 560 shows the delayed complementary differential signal -g (tr) generated by the differential amplifier 114 and the delay element 110 that can produce a 0.6T delay. The optical intensity graph 562 responds to the difference. The optical power I (t) of the optical signal after the motion signal g⑴ and the delayed complementary differential signal -g (t-r) is modulated. Similar to the optical power waveform I⑴ produced by the transmitter 100 illustrated in the optical intensity pattern 318 shown in FIG. 3, the optical power waveform I⑴ displayed by the optical intensity pattern 562 is also a three-level signal, which has Smaller peaks centered at the allelic boundary. In addition, the representative bit sequence and signal shown in FIG. 8 includes a received signal pattern 564 for implying the received signal s (t). The representative system uses a 150km long, 16ps / (km.nm) After the fiber with dispersion coefficient is propagated, the analog electrical output is received in a noise-free detector. Compared with the signal encoded by the boundary signal processing method obtained by the method in the prior art with T delay, the received signal in FIG. 8 shows a significant improvement in dispersion tolerance. FIG. 9A shows an analog 10 Gb / s optical eye diagram 600 of a received signal. The signal is based on a delay element 110 in FIG. 7A having a delay (100 ps delay) equal to one full bit period. The transmitter 500 uses a 150 km long, 16 ps / (km.nm) dispersion fiber for transmission. Compared with the NRZ transmitter known in the prior art, the transmitter 500 has improved dispersion tolerance. However, the optical eye diagram 600 shows a relatively closed shape, indicating that the dispersion of optical signals using a 150 km long, 16 ps / (km.nm) dispersion fiber has a poor tolerance for dispersion. Poor dispersion tolerance is manifested in inter-symbol interference, and these interferences are not caused by the accumulation of optical noise generated by various components (such as fiber amplifiers), because these effects are not covered in the simulation diagram . Reducing the delay time generated by the delay element 110 to less than one full bit period improves the dispersion tolerance of the transmitter 500 of FIG. 7A. In this way, a peak chirp with a maximum chirp can be generated at the bit boundary in the optical power waveform 562 shown in FIG. 8. The advantage of these peak chirps is that the interference that occurs when the 1-bits are widened under the influence of fiber dispersion can be used to preserve the continuous 1-bit amplitude. FIG. 9B shows an analog 10Gb / s optical eye diagram 602 of a received signal. The signal is a dispersion-tolerant transmitter 500 with a delay element 110 that can generate a 0.6T delay (60ps delay) in FIG. 7A. Long, 16ps / (km.nm) dispersion fiber for transmission. Compared with the transmitter 500 using the delay time of a complete bit period in FIG. 7A, the dispersion-tolerant transmitter 500 using a delay of less than one bit period will produce a larger amount in the light 200302934 academic eye diagram 602 Opening. The larger optical eye opening represents a system that has improved tolerance to dispersion, which reduces the bit error rate at the receiver. The dispersion tolerance of the transmitter 100 in FIG. 1 is greater than the dispersion tolerance of the transmitter 500 in FIG. 7A. The reason why the dispersion tolerance of the transmitter 500 in Fig. 7A is relatively low is that the signal of equal amplitude and opposite sign is not used to drive the modulator 130. This will cause some frequency sound in the optical signal, which will adversely affect the propagation. FIG. 10 shows another embodiment of a dispersion-tolerant transmitter 650 according to the present invention, which includes a parallel data input. The parallel input dispersion tolerant transmitter 650 is similar to the serial input dispersion tolerant transmitter 500 described with reference to FIG. 7A. However, as described in conjunction with FIG. 5, the transmitter 650 includes a parallel dual binary precoder 402 and a time division multiplexer 404. The parallel double-binary pre-encoder 402 includes a parallel input data bus 406 having a width equal to w; and the data input is f to. The parallel bi-binary pre-encoder 402 encodes the data on the data bus 406 in a parallel manner. The parallel encoding rate is equal to 1 / w times the bit rate of the equivalent serial data stream. The w-wide output 408 of the parallel bi-binary pre-encoder is coupled to a time division multiplexer 404. The time-division multiplexer 404 generates a binary precoding bit sequence m on the output 410 and generates a complementary binary precoding bit sequence 5 on the complementary output 412 in the form of serial data. The bit rate is w times faster than the data input V to d (w ~) of the data input sent to the parallel bi-binary precoder 200302934 402. The output 410 of the time division multiplexer 404 is complementary to The output 412 is connected to a differential amplifier 114. In another embodiment, a single-input amplifier (not shown) is connected to the output 410 and the complementary output 412 of the time division multiplexer 404. In other embodiments, the differential amplifier 114 is not used, and the time division multiplexer 404 will generate a signal with an output voltage sufficient to directly drive a modulator. The differential amplifier Π4 can convert the binary The encoded data signal and the complementary binary pre-encoded data signal are respectively converted into a differential signal g⑴ above the differential output 124 and a complementary differential signal -g⑴ above the complementary differential output 126. The differential signals g⑴ and Complementary differential The signal -g⑴ is a two-level data signal with limited bandwidth. In one embodiment, there is a bias network (such as a T-shaped bias source 128) for adding a bias voltage to the differential signal. g⑴ and one of the complementary differential signals -g⑴. In the embodiment shown in the figure, the T-shaped bias source 128 adds a bias voltage to the complementary differential signal _g⑴ to generate a Complementary differential signal containing DC offset voltage -g (t) + Vbias. In another embodiment, 'the differential amplifier 114 includes a bias voltage source for adding a bias voltage to the differential signal g⑴ and one of the complementary differential signals -g⑴. The transmitter 650 includes a delay element 110, which has an input 112 coupled to a complementary differential output 126 of the differential amplifier 114. In the other- In the embodiment, the delay element 110 is coupled to the differential output 124 of the differential amplifier 114. In another embodiment, a second delay element 42 200302934 (not shown in the figure) is coupled to the The differential output 124 of the differential amplifier Π4. The delay element 110 will The complementary differential signal -g (t) + Vbias of the DC offset voltage is delayed relative to the differential signal g⑴ for a period of time r to generate a delayed complementary differential signal -g (tr) + Vbias containing the DC offset In one embodiment, r represents a time corresponding to less than one bit period of the binary pre-encoded data signal. In one embodiment, the range of r is between 0.4T and 0.8T. As described in conjunction with FIG. 5, in order to produce the highest dispersion tolerance, the optimal delay generated by the delay element 110 is a function of the bandwidth of the signals g⑴ and -g⑴. For example, when these When the signal bandwidth is 75% (75%) of the bit rate, the optimal delay time is about 0.6T. In addition, the transmitter 650 includes an optical data modulator 130. In the embodiment shown in the figure, the optical data modulator 130 is a differential input MZI data modulator. The first data input 132 of the differential input data modulator 130 is coupled to the differential output 124 of the differential amplifier 114; the second data input 134 of the differential input data modulator 130 is coupled to An output 120 of the delay element 110. In another embodiment, the delay element 110 is coupled to the output 124 of the differential amplifier Π4. In this embodiment, the first data input 132 of the differential input data modulator 130 is coupled to the output 120 of the delay element 110, and the second data input 134 of the differential input data modulator 130 is It is coupled to a complementary differential output 126 of the differential amplifier 114. The differential input data modulator 130 also includes an optical input 136 for receiving an optical signal from a light source (e.g., a laser 138). The data modulator 43 200302934 130 can modulate a continuous wave optical signal in response to the differential signal g⑴ and the delayed complementary differential signal containing a DC offset voltage -g (t-r) + Vbias. In one embodiment, after the bias voltage Vbias is adjusted, when g (t) + g (t- r) -Vbias is equal to g (t) + g (t- r) -Vbias maximum 値 and g⑴ + g When (t-τ) -Vbias is the minimum 値 average 光学, the optical output power output by the differential input data modulator no will be at the minimum power level. Although the present invention has been particularly shown and described with reference to the preferred embodiments in detail, those skilled in the art should understand that without departing from the spirit and scope of the present invention as defined herein, Make various modifications in form and detail. [Brief description of the drawings] (I) Schematic part The advantages of the present invention can be understood with the drawings. In different drawings, the same component symbols represent the same structural elements and features. This drawing does not need to be proportionally scaled, but will be enlarged when explaining the principle of the present invention. Figure 1 shows a specific embodiment of a dispersion-tolerant transmitter according to the present invention, which includes a series of data inputs. Fig. 2A shows the transmitted optical amplitude conversion characteristics of a MZI data modulator that can be used in the dispersion-tolerant transmitter of the present invention. Figure 2B shows the transmitted optical intensity conversion characteristics of a MZI data modulator that can be used in the dispersion-tolerant transmitter of the present invention. Figure 3 shows a representative 200302934 bit sequence and signal for the dispersion-tolerant transmitter operation in Figure 1. FIG. 4A shows an analog 10Gb / s optical eye diagram of a received signal. The signal is transmitted with a delay element in FIG. 1 having a delay (100 ps delay) equal to one full bit period. It uses a 150km long, 16ps / (km.nm) dispersion fiber for transmission. FIG. 4B shows an analog 10Gb / s optical eye diagram of a received signal. The signal is a dispersion-tolerant transmitter with a delay element capable of generating a 0.75T delay (75ps delay) in FIG. /(km.nm) dispersion fiber. Figure 5 shows a specific embodiment of a dispersion-tolerant transmitter according to the present invention, which includes a parallel data input. FIG. 6 is a functional block diagram of a specific embodiment of a four-bit-wide (w = 4) parallel dual binary pre-encoder circuit, which can be used with the parallel-input dispersion-tolerant transmitter of FIG. 5. FIG. 7A shows another embodiment of a dispersion-tolerant transmitter according to the present invention, which includes a series of data inputs. FIG. 7B shows another embodiment of a dispersion-tolerant transmitter according to the present invention, which includes a series of data inputs. Figure 8 shows a representative bit sequence and signal for the dispersion-tolerant transmitter operation in Figure 7A. FIG. 9A shows an analog 10 Gb / s optical eye diagram of a received signal. The signal is based on the transmitter in FIG. , Using 150km long, 16ps / (km.nm) dispersion fiber for transmission. 45 200302934 Figure 9B shows an analog 10Gb / s optical eye diagram of the received signal. This signal is a transmitter with a delay element that can generate a 0.6T delay (60ps delay) as shown in Figure 8. It uses a 150km long, 16ps / ( km.nm) dispersion fiber. FIG. 10 shows another embodiment of a dispersion-tolerant transmitter according to the present invention, which includes a parallel data input. (II) Symbols for component representation 100,500,550 Dispersion-tolerant transmitter including serial data input 102,552 Serial dual binary pre-encoder 104 Serial dual binary pre-encoder 102,552 Serial data input 106 Serial dual-binary pre Set output of encoder 102 108 Complementary output of tandem binary pre-encoder 102 110 Delay element 112 Input of delay element 110 Differential amplifier 118 First input of differential amplifier 120 Output of delay element Π0 122 Differential Second input of amplifier 114 124 Differential output of differential amplifier Π4 Complementary differential output of differential amplifier 114 128 T-shaped bias source 130 Optical data modulator 132 First data input of optical data modulator 130 200302934 134 135 136 137 138 400,650 402 404 406 408 410 412 450 452,468 454 456 457 458 460 462 464 466 470 Second data input filter (first filter) of the optical data modulator 130 Optical input of the data modulator 130 Second filter laser includes dispersion-tolerant transmitter with parallel data input Parallel bi-binary pre-encoder The time-division multiplexer tolerates the parallel data input of the dispersive transmitter 400. The output of the parallel bi-binary pre-encoder 402. The output of the time-division multiplexer 404. The complementary output of the time-division multiplexer 404. The data input of the D-type flip-flop flip-flop 452 is set to the encoder circuit. The clock input of the flip-flop 452 shares the complementary output of the clock flip-flop 452. One of the inputs or the gate 462 is mutually exclusive. The data input of the gate 462 output flip-flop 468 is mutually exclusive or the output of the other input of the gate 462 flip-flop 468 47 472 200302934 474 476 554 556 buffer parallel dual binary pre-encoder circuit 450 data output serial double Output of Binary Precoder 552 Tandem Complementary Output of Dual Binary Precoder 552
4848