JP3657209B2

JP3657209B2 - Anomaly detection circuit for optical receiver and optical receiver

Info

Publication number: JP3657209B2
Application number: JP2001235929A
Authority: JP
Inventors: 錬太郎吉越
Original assignee: Mitsubishi Electric Corp
Current assignee: Mitsubishi Electric Corp
Priority date: 2000-10-26
Filing date: 2001-08-03
Publication date: 2005-06-08
Anticipated expiration: 2021-08-03
Also published as: JP2002204208A

Description

【０００１】
【発明の属する技術分野】
本発明は、光通信等に用いられ、光信号から電気信号に光電変換する光電変換素子を備えた光受信器において信号の異常を検出する光受信器用異常検出回路に関し、特に、符号誤り率に応じた異常検出に関する。
【０００２】
【従来の技術】
図９は従来の光受信器用異常検出回路２１のブロック図である。図９において、１は入力された光信号を電流信号に変換する光電変換素子、２は電流信号を適切な振幅の電圧信号に変換するトランスインピーダンスプリアンプ、３はトランスインピーダンスプリアンプ２のゲインを設定する帰還抵抗、４は前記トランスインピーダンスプリアンプ２の出力信号を増幅するポストアンプ、１７はポストアンプ４の出力信号の振幅に比例する信号を出力するピーク検波回路、１８はアラーム信号発出の閾値信号を出力する閾値生成回路、１９はピーク検波回路１７の出力信号が閾値生成回路１８の出力信号レベルより高い（または低い）場合にアラーム信号を出力する比較器、１０はアラーム信号出力端子である。
【０００３】
次にこの光受信器用異常検出回路２１の動作について説明する。図９に示す光受信器用異常検出回路２１において、強い強度（振幅）の光信号を「Ｈ」、弱い強度の光信号を「Ｌ」とするデジタル光信号が光電変換素子１に入力される。光電変換素子１はこの光信号を電流信号に変換する。この電流信号はトランスインピーダンスプリアンプ２に入力され、トランスインピーダンスプリアンプ２は、この電流信号を入力電流信号振幅に応じて適切な振幅の電圧信号に変換する。トランスインピーダンスプリアンプ２において電流信号から変換された電圧信号は、ポストアンプ４において、ピーク検波回路１７での振幅検出機能が正常に動作するように増幅される。ピーク検波回路１７は、ポストアンプ４の出力信号の振幅に比例した信号を出力する。また閾値生成回路１８は、用途に応じたレベルの閾値信号を生成する。比較器１９は、ピーク検波回路１７の出力信号と閾値生成回路１８の閾値信号とを比較し、例えばピーク検波回路１７の出力信号が閾値生成回路１８の閾値信号より低くなった場合、アラーム信号を出力し、このアラーム信号が、アラーム信号出力端子１０からこの光受信器用異常検出回路２１のアラーム出力信号として出力される。このような構成により、光電変換素子１に入力される光入力強度が所定値を下回った場合に、アラーム信号出力端子１０よりアラーム信号を出力することができる。
【０００４】
【発明が解決しようとする課題】
昨今、光受信器において、光信号の符号誤り率が所定の比率以上となった場合にアラームを発生させる機能を備えたものが望まれているが、符号誤り率は信号のノイズ成分に大きく依存するため、ノイズ成分を考慮しない光入力強度に基づく上記従来の光受信器用の異常検出回路を、このような符号誤り率の増加に応じた異常を検出する回路として利用することができなかった。
【０００５】
本発明はこのような課題に鑑みてなされたものであり、その目的は、光受信器において、光電変換素子等の構成部品の個体バラツキ，温度，または電圧の変動等により信号中のノイズ成分が増大し符号誤り率が増大した場合にこれを検出することにある。
【０００６】
【課題を解決するための手段】
本発明にかかる光受信器用異常検出回路は、光電変換素子により光信号から光電変換された電気信号の強度と、所定の識別閾値と、を比較する主識別回路と、前記電気信号の強度と、前記所定の識別閾値と異なる参照閾値と、を比較する参照識別回路と、前記主識別回路および前記参照識別回路の比較結果に基づいて信号の異常を検出する演算回路と、を備え、前記演算回路は、電気信号の強度の前記識別閾値に対する大小の識別結果と前記参照閾値に対する大小の識別結果とが異なる場合の回数を所定期間積算し、その積算した回数が、与えられた符号誤り率に応じて決定される前記主識別回路および前記参照識別回路の各比較結果に差異が発生する確率と前記所定期間の間に伝送する全符号数とから算出される異常検出用閾値を超えた場合を信号の異常として検出するものである。
【０００８】
また本発明では、前記識別閾値は、主識別回路においてその識別閾値を設定したときに、識別閾値より小さくなるべき信号が識別閾値より大きい信号として識別される確率と、識別閾値より大きくなるべき信号が識別閾値より小さい信号として識別される確率と、が等しくなるように、設定されるのが好適である。
【０００９】
また本発明にかかる光受信器用異常検出回路は、所定の識別タイミングで、光電変換素子により光信号から光電変換された電気信号の強度と、所定の識別閾値と、を比較する主識別回路と、前記所定の識別タイミングと異なる参照タイミングで、前記電気信号の強度と、所定の識別閾値と、を比較する参照識別回路と、前記主識別回路および参照識別回路の識別結果に基づいて信号の異常を検出する演算回路と、を備え、前記演算回路は、電気信号の強度の前記識別タイミングにおける前記識別閾値に対する大小の識別結果と前記参照タイミングにおける前記識別閾値に対する大小の識別結果とが異なる場合の回数を所定期間積算し、その積算した回数が、与えられた符号誤り率に応じて決定される前記主識別回路および前記参照識別回路の各比較結果に差異が発生する確率と前記所定期間の間に伝送する全符号数とから算出される異常検出用閾値を超えた場合を信号の異常として検出するものである。
【００１１】
また本発明では、前記識別タイミングは、主識別回路においてその識別タイミングを設定したときに、アイにおいて早い側の信号変化タイミングで前記識別閾値となるべき電気信号が実際にはその識別タイミングより遅いタイミングで識別閾値となる確率と、アイにおいて遅い側の信号変化タイミングで前記識別閾値となるべき電気信号が実際にはその識別タイミングより早いタイミングで識別閾値となる確率と、が等しくなるように、設定されるのが好適である。
【００１２】
また本発明にかかる光受信器用異常検出回路は、所定の識別タイミングで、光電変換素子により光信号から光電変換された電気信号の強度と、所定の識別閾値と、を比較する主識別回路と、所定の識別タイミングで、前記電気信号の強度と、前記所定の識別閾値と異なる参照閾値と、を比較する第一の参照識別回路と、前記所定の識別タイミングと異なる参照タイミングで、前記電気信号の強度と、前記所定の識別閾値とを比較する第二の参照識別回路と、前記主識別回路、第一の参照識別回路および第二の参照識別回路の識別結果に基づいて信号の異常を検出する演算回路と、を備え、前記演算回路は、電気信号の強度の前記識別閾値に対する大小の識別結果と前記参照閾値に対する大小の識別結果とが異なる場合の回数を所定期間積算し、その積算した回数が、与えられた符号誤り率に応じて決定される前記主識別回路および前記参照識別回路の各比較結果に差異が発生する確率と前記所定期間の間に伝送する全符号数とから算出される第１の異常検出用閾値を超えた場合、又は電気信号の強度の前記識別タイミングにおける前記識別閾値に対する大小の識別結果と前記参照タイミングにおける前記識別閾値に対する大小の識別結果とが異なる場合の回数を所定期間積算し、その積算した回数が、与えられた符号誤り率に応じて決定される前記主識別回路および前記参照識別回路の各比較結果に差異が発生する確率と前記所定期間の間に伝送する全符号数とから算出される第２の異常検出用閾値を超えた場合の少なくともいずれか一方が発生した場合を信号の異常として検出するものである。
【００１３】
また本発明にかかる光受信器は、上記いずれかの光受信器用異常検出回路を備える。
【００１４】
【発明の実施の形態】
図１は本発明の実施の形態にかかる光受信器用異常検出回路２０のブロック図である。図１において、光受信器用異常検出回路２０は、入力された光信号を電流信号に変換する光電変換素子１（例えばフォトダイオード）、電流信号を適切な振幅の電圧信号に変換するトランスインピーダンスプリアンプ２、トランスインピーダンスプリアンプ２の増幅率（すなわちゲイン）を設定する帰還抵抗３、トランスインピーダンスプリアンプ２の出力信号を増幅するポストアンプ４、ポストアンプ４の出力信号からクロック信号を抽出するクロック抽出回路５、ポストアンプ４の出力信号レベル（振幅）をそれぞれ適切にシフトするレベルシフト回路６（６ａ，６ｂ，６ｃ，・・・）、各レベルシフト回路６毎に設けられ、クロック抽出回路５の出力クロック信号をそれぞれ適切に遅延させるクロック遅延回路７（７ａ，７ｂ，７ｃ，・・・）、各レベルシフト回路６およびクロック遅延回路７毎に設けられ、これらレベルシフト回路６およびクロック遅延回路７の出力信号に基づいて信号のレベルが「Ｈ」であるか「Ｌ」であるかを識別する識別回路８（８ａ，８ｂ，８ｃ）、各識別回路８の識別結果の出力に基づいて、これら識別結果のカウントおよび演算を行いアラーム信号を発出する演算回路９、および、演算回路９の発出したアラーム信号を出力するアラーム出力端子１０を備える。
【００１５】
次にこの異常検出回路２０の動作について説明する。例えば一定のパルス間隔でランダムに強度変調されたデジタル光パルス信号が光電変換素子１に入力され、ここで電気信号としての電流パルス信号に変換される。その電流パルス信号は、トランスインピーダンスプリアンプ２、およびポストアンプ４により、適切な振幅に増幅される。
【００１６】
図２に、増幅後の電流パルス信号のアイ（ｅｙｅ）パターンを示す。この図において、横軸は時間、また縦軸は振幅（レベル）を示す。当該アイパターンにおいて、信号中の符号のレベル（すなわち振幅）が「Ｈ」（：大）であるか「Ｌ」（：小）であるかを識別するために、アイ２２の中心部分（○印）に真の識別点１１を設定する。各符号毎に真の識別点１１とその信号レベルとが比較され、例えば、信号レベルが真の識別点１１の識別閾値以上のレベルであればその符号は「Ｈ」（：大）と識別され、また真の識別点１１の識別閾値より小さいレベルであればその符号は「Ｌ」（：小）と識別される。
【００１７】
この真の識別点１１での「Ｈ」あるいは「Ｌ」の識別は主識別回路８ａで行われる。この識別を精度よく行うため、信号は、該主識別回路８ａに入力される前に、その振幅および時間がそれぞれオフセットされる。振幅のオフセットは第一のレベルシフト回路６ａにより行われ、また時間のオフセットは第一のクロック遅延回路７ａにより行われる。すなわち、第一のレベルシフト回路６ａにより、真の識別点１１が所定のアイパターンの縦方向（振幅方向）のほぼ中央となるように調整され、また第一のクロック遅延回路７ａにより、真の識別点１１が同じく所定のアイパターンの横方向（時間軸方向）のほぼ中央となるように調整される。これらオフセットにより、主識別回路８ａにおける識別閾値および識別タイミングが設定あるいは調整される。なお、本実施の形態では、第一のレベルシフト回路６ａはポストアンプ４の出力および主識別回路８ａの入力に接続され、また第一のクロック遅延回路７ａはクロック抽出回路５の出力および主識別回路８ａの入力にそれぞれ接続される。
【００１８】
また、本発明では、真の識別点１１の他に参照識別点１２が一つ又は複数設けられる。この参照識別点１２での「Ｈ」あるいは「Ｌ」の識別は参照識別回路（８ｂ，８ｃ，・・・）で行われる。この識別は、第１のレベルシフト回路６ａ、第１のクロック遅延回路７ａ、および主識別回路８ａと並列に、また同じ接続関係で、レベルシフト回路（６ｂ，６ｃ，・・・）、クロック遅延回路（７ｂ，７ｃ，・・・）、および参照識別回路（８ｂ，８ｃ，・・・）を設け、当該レベルシフト回路（６ｂ，６ｃ，・・・）のレベルシフト量、およびクロック遅延回路（７ｂ，７ｃ，・・・）のクロック遅延量を、それぞれ第１のレベルシフト回路６ａのレベルシフト量（振幅オフセット量）、および第１のクロック遅延回路７ａのクロック遅延量（時間オフセット量）と異なる量とすることにより実現される。この参照識別点１２（図２中×印）は、図２に示すアイパターン上では、真の識別点１１（図２中○印）と異なる位置の点として示すことができる。各参照識別点１２は、このアイパターン上では、真の識別点１１とのレベルシフト量あるいはクロック遅延量との差異に応じて、この真の識別点１１から離間した位置に設定されることになる。すなわち、各参照識別回路（８ｂ，８ｃ，・・・）は、それぞれ識別閾値または識別タイミングとは異なる参照閾値または参照タイミングにより二値の識別（すなわち、それぞれに設定された閾値に対する大小の識別）を行う。
【００１９】
各識別回路８で得られた識別結果は、それぞれ演算回路９に入力される。演算回路９はカウンタ機能を有し、所定期間内で真の識別点１１の識別回路の識別結果と参照識別点１２の識別回路の識別結果に差異があった場合の数をカウントし、当該所定期間内に伝送される全符号のビット数に対する識別結果の異なる場合のカウント数の割合を求める。符号識別結果とアイパターン上の識別点の位置には相関があるため、真の識別点１１と参照識別点１２での識別結果が異なる場合の割合と、それぞれの識別点のアイパターン上での位置関係から、真の識別点１１における符号誤り率を推定することができる。すなわち識別点の位置関係とカウント測定期間と識別結果の差異のカウント数量をそれぞれ適切に設定することにより、演算回路９から符号誤り率が所定値を超えたときにアラーム信号を発出することが可能となる。
【００２０】
実施の形態１．
本発明の第一の実施の形態として、真の識別点１１に対してそのレベルが異なる（例えば高い）点１点を参照識別点１２とした場合について説明する。図３にこの場合のアイパターンを示す。図に示すように、ここでは、一例として、真の識別点１１のレベル（すなわち識別閾値）を「Ｈ」信号のレベルより２ａ低く、また参照識別点１２のレベル（すなわち参照閾値）を「Ｈ」信号のレベルよりａ低く設定している。以下、この場合の、真の識別点１１での符号誤り率と、カウント測定期間と、真の識別点１１と参照識別点１２での識別結果の異なる場合の数について説明する。なお、以下では、信号伝送速度を２．５ギガビット／秒（Ｇｂ／ｓ）、演算回路９による差異のカウント期間（後述）を４０マイクロ秒（μｓ）とし、符号誤り率が１０^−３（０．００１）を越えた際にアラームを発出させるような構成とする場合について説明する。
【００２１】
図３のアイパターンでは、各信号を一本のラインで表しているが、信号に含まれるノイズ成分を考慮すると、「Ｈ」信号および「Ｌ」信号のレベルは、それぞれ正規分布に基づいた確率でばらついて広がると仮定することができる。これを表したのが図４である。この図では、縦軸はレベル、横軸は確率を示し、上段の曲線が「Ｈ」信号の検出レベルの確率分布を、また下段の曲線が「Ｌ」信号の検出レベルの確率分布をそれぞれ示す。この図において、「Ｌ」信号を「Ｈ」信号と誤る確率（すなわち、「Ｌ」信号となるべき信号が「Ｈ」信号として識別される確率）は、確率分布範囲１３における確率分布の積分値（すなわち図４では領域１３の面積）となり、他方「Ｈ」信号を「Ｌ」信号と誤る確率（すなわち、「Ｈ」信号となるべき信号が「Ｌ」信号として識別される確率）は確率分布範囲１４における確率分布の積分値（すなわち図４では領域１４の面積）となる。つまり、正規分布において、確率分布範囲１３の面積と確率分布範囲１４の面積との和が、真の識別点１１における符号誤り率Ｐ（Ｒ）となる。ここで、本実施の形態では、確率分布範囲１３および１４の面積がほぼ等しくなるように、真の識別点１１を設定する。すなわち換言すれば、識別閾値以上で「Ｌ」信号が「Ｈ」信号と誤認される確率（すなわち領域１３の面積）と、識別閾値以下で「Ｈ」信号が「Ｌ」信号と誤認される確率（すなわち領域１４の面積）と、がほぼ等しくなるように、真の識別点１１のレベル（すなわち識別閾値）を設定する。
【００２２】
参照識別点１２における識別結果と真の識別点１１における識別結果とが異なる確率Ｐ（Ｄ）は、図４に示されるように、確率分布範囲１５と確率分布範囲１３の和になる。ここで、参照識別点１２のレベル（すなわち参照閾値）において二値のうちの一方の信号（この場合「Ｌ」信号）を他方の信号（この場合「Ｈ」信号）と識別する確率は非常に低いためこれをゼロと仮定すると、確率分布範囲１３と確率分布範囲１４がほぼ等しいことから、前記Ｐ（Ｄ）は確率分布範囲１５と確率分布範囲１４の和であるといえる。これは、参照識別点１２における「Ｈ」信号の正規分布の上側確率（すなわちこの場合、参照閾値以下のレベルでその信号が「Ｈ」信号である確率）に相当する。ここで、符号誤り率Ｐ（Ｒ）が１０^−３（０．００１）の時を考える。この場合、確率分布範囲１３と確率分布範囲１４の和が１０^−３であるので、各確率分布範囲１３，１４（すなわち正規分布における上側確率）はそれぞれ５×１０^−４（０．０００５）である。正規分布において上側確率が５×１０^−４であるときは、その分散値をσとすると、真の識別点１１のレベル２ａ＝３．３σのときである。この時、参照識別点１２のレベルａではその半分の１．６５σとなり、この時の正規分布の上側確率はおよそ０．０４９５となる。したがって、符号誤り率が１０^−３の時、真の識別点１１での識別結果と、参照識別点１２での識別結果に差異が発生する確率（差異発生確率）は０．０４９５であるといえる。演算回路９においてカウンタを動作させる所定期間（例えば４０マイクロ秒）の間に伝送する全符号数は、伝送速度が２．５Ｇｂ／ｓの場合、１０万ビットである。したがって、演算回路９のカウンタにより識別結果の差異をカウントしてこれを所定期間（４０マイクロ秒間）積算し、その結果がこの期間における前記差異発生確率０．０４９５に相当する数、すなわち１０万×０．０４９５＝４９５０カウントを超えた場合を異常としてアラームを発出させれば、これを、符号誤り率１０^−３を超えた場合の異常検出およびアラーム発出とすることができる。
【００２３】
実施の形態２．
本実施の形態では、図５に示すように、真の識別点１１とタイミングの異なる（例えば早い）点を参照識別点１２とし、同じ識別閾値に対する大小の識別結果の差異の発生回数をカウントすることにより、アラーム発出を行う。以下、この原理について説明する。
【００２４】
アイ（ｅｙｅ）の両側の、「Ｈ」→「Ｌ」または「Ｌ」→「Ｈ」に信号が変化するタイミング（すなわち図５および図６のタイミングＡ，Ｂ）において、識別閾値となるべき信号が実際に識別閾値となるタイミングの確率分布は、実施の形態１．と同様に、通常、図６のような正規分布となる。
【００２５】
本実施の形態では、タイミングＡで識別閾値となるべき信号が実際にはタイミングＣより遅いタイミングで識別閾値となる確率（図６の３４）と、タイミングＢで識別閾値となるべき信号が実際にはタイミングＣより早いタイミングで識別閾値となる確率（図６の３３）と、が等しくなるタイミングＣが、識別タイミングに設定される。また、参照識別点１２の参照タイミングＤは、識別タイミングＣと異なる（例えば早い）タイミングであって、タイミングＢで識別閾値となるべき信号が実際にそのタイミングで識別閾値となる確率のほとんど無いタイミング（すなわち、図６のＢに対応する確率分布から外れたタイミング）、に設定される。なお、真の識別点１１および参照識別点１２の識別閾値は、例えば、「Ｈ」レベルと「Ｌ」レベルとのちょうど中間の値とする。
【００２６】
このように二つの識別点が設定されている状況下で、信号が図６に示す正規分布どおりに出現した場合には、タイミングＡで識別閾値となるべき信号が参照タイミングＤより遅いタイミングで識別閾値となる実際の確率（図６の３６の面積）は、参照タイミングＤと識別タイミングＣとの間のタイミングで信号が識別閾値となる正規分布上の確率（図６の３５と３３との面積の合計）に等しくなる。しかしながら、符号誤り率（正規分布上では図６の３３と３４との面積の合計に相当する）が増大すると、その分、参照タイミングＤと識別タイミングＣとの間のタイミングで信号が識別閾値となる実際の確率は、タイミングＡで識別閾値となるべき信号が参照タイミングＤより遅いタイミングで識別閾値となる正規分布上の確率（図６の３６の面積）より大きくなる。
【００２７】
ここで、図７に示すように、信号が参照タイミングＤでは「Ｈ」レベルであって、かつ識別タイミングＣでは「Ｌ」レベルである場合、および、信号が参照タイミングＤでは「Ｌ」レベルであって、かつ識別タイミングＣでは「Ｈ」レベルである場合には、信号は、参照タイミングＤと識別タイミングＣとの間で識別閾値となる。すなわち、参照識別点１１における識別閾値に対する識別結果と、真の識別点１２における識別閾値に対する識別結果と、が異なる場合には、信号は、識別タイミングＣと参照タイミングＤとの間のタイミングで識別閾値となる。
【００２８】
したがって、タイミングＡで「Ｈ」→「Ｌ」または「Ｌ」→「Ｈ」に変化する全信号の数のうち、二つの識別点１１，１２で異なる識別結果となる信号の数の比率が、タイミングＡで識別閾値となるべき信号が参照タイミングＤより遅いタイミングで識別閾値となる正規分布上の確率（図６の３６の面積）より高くなった場合は、正規分布上想定される符号誤り率より誤り率の高い状態であると識別することができる。
【００２９】
ここで、符号誤り率が１０^−３を超えた場合にアラームが発出される例について考える。正規分布上での符号誤り率が１０^−３である場合、タイミングＡで識別閾値となるべき信号がタイミングＣより遅いタイミングで識別閾値となる確率（図６の３４）と、タイミングＢで識別閾値となるべき信号がタイミングＣより早いタイミングで識別閾値となる確率（図６の３３）とは、いずれも５×１０^−４となる。正規分布において上側確率が５×１０^−４となるのは、３．３σ（σは分散値）のときである。図５のように識別タイミングＣ（２ｂ）に対して参照タイミングＤ（ｂ）を設定すると、参照タイミングＤ（ｂ＝１．６５σ）より上側の確率（図６の３６の面積）は、０．０４９５となる。タイミングＡの前後における信号の推移の全パターン（「Ｈ」→「Ｈ」，「Ｈ」→「Ｌ」，「Ｌ」→「Ｈ」，「Ｌ」→「Ｌ」）のうち、「Ｈ」→「Ｌ」または「Ｌ」→「Ｈ」に値が変化する確率はその１／２であるから、信号が正規分布にしたがっている場合には、全信号のうち二つの識別点１１，１２で識別結果が異なる場合の確率（図６の３３と３４との面積の合計）は、０．０２４７５（＝０．０４９５／２）である。これは、実施の形態１の場合の１／２に相当する。演算回路９のカウンタでカウンタを動作させる所定期間（例えば４０マイクロ秒）の間に伝送する全符号数は、伝送速度が２．５Ｇｂ／ｓの場合、１０万ビットである。したがって、演算回路９のカウンタにより識別結果の差異をカウントしてこれを所定期間（例えば４０マイクロ秒）積算し、その結果が、この期間における差異発生確率０．０２４７５に相当する数、すなわち１０万×０．０２４７５＝２４７５カウントを超えた場合に、符号誤り率１０^−３を超えた異常としてアラームを発出させればよい。なお、本実施形態でも、真の識別点１１の識別は主識別回路８ａにより行い、参照識別点１２の識別は参照識別回路８ｂにより行えばよい。
【００３０】
実施の形態３．
図８に示すように、本発明の第３の実施の形態として、真の識別点１１に対してそのレベルの異なる（例えば高い）点、およびそのタイミングの異なる（例えば早い）点の双方を参照識別点１２として用い、前の実施の形態１および２と全く同様の手法で当該参照識別点１２と真の識別点１１との差異発生回数をカウントすることにより、所定の符号誤り率を越えた場合のアラーム発出を行うことができる。この場合は実施の形態１と実施の形態２とを組み合わせ、レベル方向および時間軸方向の各々について識別結果の差異をカウントし、両方のカウンタが所定の回数を超えた場合、もしくは、いずれか一方のカウンタが所定の回数を超えた場合にアラームを発出することにより、所定の符号誤り率を越えた場合のアラーム発出を実現することができる。
【００３１】
【発明の効果】
以上説明したように、本発明によれば、識別閾値に対する識別結果と、識別閾値と異なる参照閾値に対する識別結果とに基づいて符号誤り率に応じた光信号の異常を検出する回路を実現できる。
【００３２】
また本発明によれば、さらに、電気信号の強度の前記識別閾値に対する大小の識別結果と、参照閾値に対する大小の識別結果と、が異なる場合の回数を、所定期間積算し、積算した回数が所定値を超えた場合を異常とすることで、符号誤り率に応じた光信号の異常を検出する回路を実現できる。
【００３３】
また本発明によれば、さらに、識別閾値を、主識別回路においてその識別閾値を設定したときに、識別閾値より小さくなるべき信号が識別閾値より大きい信号として識別される確率と、識別閾値より大きくなるべき信号が識別閾値より小さい信号として識別される確率と、が等しくなるように設定することで、符号誤り率に応じた光信号の異常を検出する回路を実現できる。
【００３４】
また本発明によれば、識別タイミングにおける識別結果と、識別タイミングと異なる参照タイミングにおける識別結果とに基づいて、符号誤り率に応じた光信号の異常を検出する回路を実現できる。
【００３５】
また本発明によれば、さらに、電気信号の強度の前記識別タイミングにおける前記識別閾値に対する大小の識別結果と、前記参照タイミングにおける前記識別閾値に対する大小の識別結果と、が異なる場合の回数を、所定期間積算し、積算した回数が所定値を超えた場合を異常とすることで、符号誤り率に応じた光信号の異常を検出する回路を実現できる。
【００３６】
また本発明によれば、さらに、識別タイミングを、主識別回路においてその識別タイミングを設定したときに、アイにおいて早い側の信号変化タイミングで識別閾値となるべき電気信号が実際にはその識別タイミングより遅いタイミングで識別閾値となる確率と、アイにおいて遅い側の信号変化タイミングで識別閾値となるべき電気信号が実際にはその識別タイミングより早いタイミングで識別閾値となる確率と、が等しくなるように設定することで、符号誤り率に応じた光信号の異常を検出する回路を実現できる。
【００３７】
また本発明によれば、識別タイミングにおける識別閾値に対する識別結果と、識別閾値と異なる参照閾値に対する識別結果と、識別タイミングと異なる参照タイミングにおける識別結果と、に基づいて、符号誤り率に応じた光信号の異常を検出する回路を実現できる。
【００３８】
また本発明によれば、上記した異常検出回路を備えた光受信器が実現される。
【図面の簡単な説明】
【図１】本発明の実施の形態にかかる光受信器用異常検出回路を示すブロック図である。
【図２】本発明の実施の形態にかかる光受信器用異常検出回路における信号のアイパターン、および識別点を示す図である。
【図３】本発明の第一の実施の形態にかかる光受信器用異常検出回路における入力信号のアイパターン、および識別点を示す図である。
【図４】本発明の実施の形態にかかる光受信器用異常検出回路における信号レベルの確率分布と識別点の設定を示す説明図である。
【図５】本発明の第二の実施の形態にかかる光受信器用異常検出回路における入力信号のアイパターン、および識別点を示す図である。
【図６】本発明の第二の実施形態にかかる光受信器用異常検出回路における信号値が識別閾値となるタイミングの確率分布を示す図である。
【図７】本発明の第二の実施形態にかかる光受信器用異常検出回路において、識別タイミングと参照タイミングとの間のタイミングで識別閾値となる信号を示す説明図である。
【図８】本発明の第三の実施の形態にかかる光受信器用異常検出回路における入力信号のアイパターン、および識別点を示す図である。
【図９】従来の光受信器用異常検出回路を示すブロック図である。
【符号の説明】
１光電変換素子、２トランスインピーダンスプリアンプ、３帰還抵抗、４ポストアンプ、５クロック抽出回路、６レベルシフト回路、７クロック遅延回路、８識別回路、８ａ主識別回路、８ｂ，８ｃ，・・・参照識別回路、９演算回路、１１真の識別点、１２参照識別点、２０光受信器用異常検出回路。[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an optical receiver anomaly detection circuit for detecting an anomaly of a signal in an optical receiver that is used for optical communication or the like and includes a photoelectric conversion element that photoelectrically converts an optical signal to an electrical signal. It relates to the detection of abnormalities.
[0002]
[Prior art]
FIG. 9 is a block diagram of a conventional anomaly detection circuit 21 for an optical receiver. In FIG. 9, 1 is a photoelectric conversion element that converts an input optical signal into a current signal, 2 is a transimpedance preamplifier that converts the current signal into a voltage signal having an appropriate amplitude, and 3 is a gain of the transimpedance preamplifier 2. Feedback resistor 4 is a post-amplifier for amplifying the output signal of the transimpedance preamplifier 2, 17 is a peak detection circuit for outputting a signal proportional to the amplitude of the output signal of the post-amplifier 4, and 18 is a threshold signal for issuing an alarm signal. A threshold generation circuit 19 is a comparator for outputting an alarm signal when an output signal of the peak detection circuit 17 is higher (or lower) than an output signal level of the threshold generation circuit 18, and 10 is an alarm signal output terminal.
[0003]
Next, the operation of the optical receiver abnormality detection circuit 21 will be described. In the optical receiver abnormality detection circuit 21 shown in FIG. 9, a digital optical signal having a high intensity (amplitude) optical signal “H” and a weak intensity optical signal “L” is input to the photoelectric conversion element 1. The photoelectric conversion element 1 converts this optical signal into a current signal. This current signal is input to the transimpedance preamplifier 2, and the transimpedance preamplifier 2 converts the current signal into a voltage signal having an appropriate amplitude according to the input current signal amplitude. The voltage signal converted from the current signal in the transimpedance preamplifier 2 is amplified in the postamplifier 4 so that the amplitude detection function in the peak detection circuit 17 operates normally. The peak detection circuit 17 outputs a signal proportional to the amplitude of the output signal of the post amplifier 4. The threshold generation circuit 18 generates a threshold signal having a level corresponding to the application. The comparator 19 compares the output signal of the peak detection circuit 17 and the threshold signal of the threshold generation circuit 18. For example, when the output signal of the peak detection circuit 17 becomes lower than the threshold signal of the threshold generation circuit 18, the comparator 19 This alarm signal is output from the alarm signal output terminal 10 as an alarm output signal of the optical receiver abnormality detection circuit 21. With such a configuration, an alarm signal can be output from the alarm signal output terminal 10 when the light input intensity input to the photoelectric conversion element 1 falls below a predetermined value.
[0004]
[Problems to be solved by the invention]
In recent years, optical receivers are desired to have a function for generating an alarm when the code error rate of an optical signal exceeds a predetermined ratio, but the code error rate greatly depends on the noise component of the signal. Therefore, the conventional abnormality detecting circuit for an optical receiver based on the optical input intensity that does not consider the noise component cannot be used as a circuit for detecting an abnormality according to the increase in the code error rate.
[0005]
The present invention has been made in view of such problems, and the purpose of the present invention is to reduce noise components in a signal due to individual variations of components such as photoelectric conversion elements, temperature, or voltage fluctuations in an optical receiver. This is to detect when the code error rate increases and the code error rate increases.
[0006]
[Means for Solving the Problems]
An abnormality detection circuit for an optical receiver according to the present invention includes a main identification circuit that compares the intensity of an electrical signal photoelectrically converted from an optical signal by a photoelectric conversion element with a predetermined identification threshold, the intensity of the electrical signal, A reference identification circuit that compares a reference threshold different from the predetermined identification threshold; and an arithmetic circuit that detects a signal abnormality based on a comparison result between the main identification circuit and the reference identification circuit. The arithmetic circuit integrates the number of times when the magnitude discrimination result of the electric signal strength with respect to the discrimination threshold is different from the magnitude discrimination result with respect to the reference threshold for a predetermined period, and the number of times of the integration is given code Exceeds the threshold for abnormality detection calculated from the probability that each comparison result of the main identification circuit and the reference identification circuit determined according to the error rate will differ and the total number of codes transmitted during the predetermined period Is detected as a signal abnormality .
[0008]
In the present invention, the identification threshold value is a probability that a signal that should be smaller than the identification threshold is identified as a signal larger than the identification threshold and a signal that should be larger than the identification threshold when the identification threshold value is set in the main identification circuit. Is preferably set so that the probability that the signal is identified as a signal smaller than the identification threshold is equal.
[0009]
An abnormality detection circuit for an optical receiver according to the present invention includes a main identification circuit that compares the intensity of an electric signal photoelectrically converted from an optical signal by a photoelectric conversion element with a predetermined identification threshold at a predetermined identification timing, A reference identification circuit that compares the strength of the electrical signal with a predetermined identification threshold at a reference timing different from the predetermined identification timing, and signal abnormality based on the identification results of the main identification circuit and the reference identification circuit And an arithmetic circuit for detecting The arithmetic circuit accumulates the number of times when the magnitude discrimination result with respect to the discrimination threshold at the discrimination timing of the intensity of the electric signal is different from the magnitude discrimination result with respect to the discrimination threshold at the reference timing for a predetermined period, Is calculated from the probability that a difference occurs in each comparison result between the main identification circuit and the reference identification circuit determined according to a given code error rate and the total number of codes transmitted during the predetermined period. Is detected as a signal abnormality when the detected abnormality detection threshold is exceeded .
[0011]
In the present invention, the identification timing is a timing when an electrical signal that should be the identification threshold at the earlier signal change timing in the eye is actually later than the identification timing when the identification timing is set in the main identification circuit. The probability that the discrimination threshold is equal to the probability that the electrical signal that should be the discrimination threshold at the later signal change timing in the eye actually becomes the discrimination threshold at a timing earlier than the discrimination timing is set to be equal. It is preferred that
[0012]
An abnormality detection circuit for an optical receiver according to the present invention includes a main identification circuit that compares the intensity of an electric signal photoelectrically converted from an optical signal by a photoelectric conversion element with a predetermined identification threshold at a predetermined identification timing, A first reference identification circuit that compares the intensity of the electric signal with a reference threshold different from the predetermined identification threshold at a predetermined identification timing; and a reference timing different from the predetermined identification timing; An abnormality of the signal is detected based on the identification results of the second reference identification circuit that compares the intensity with the predetermined identification threshold, and the main identification circuit, the first reference identification circuit, and the second reference identification circuit And an arithmetic circuit The arithmetic circuit integrates the number of times when the magnitude discrimination result of the electric signal strength with respect to the discrimination threshold is different from the magnitude discrimination result with respect to the reference threshold for a predetermined period, and the number of times of the integration is given code First anomaly detection calculated from a probability that a difference occurs in each comparison result of the main identification circuit and the reference identification circuit determined according to an error rate and the total number of codes transmitted during the predetermined period When the threshold value is exceeded, or the number of times when the magnitude discrimination result with respect to the discrimination threshold value at the discrimination timing of the electric signal intensity is different from the magnitude discrimination result with respect to the discrimination threshold value at the reference timing is accumulated for a predetermined period, The number of times of integration is determined according to a given code error rate, and the probability that a difference occurs in each comparison result of the main identification circuit and the reference identification circuit, and the In which at least one is detected as a signal of abnormal if it occurs in the case of exceeding the second abnormality detection threshold value is calculated from the total number of codes to be transmitted on a regular course of .
[0013]
An optical receiver according to the present invention includes any one of the above optical receiver abnormality detection circuits.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a block diagram of an optical receiver abnormality detection circuit 20 according to an embodiment of the present invention. In FIG. 1, an optical receiver abnormality detection circuit 20 includes a photoelectric conversion element 1 (for example, a photodiode) that converts an input optical signal into a current signal, and a transimpedance preamplifier 2 that converts the current signal into a voltage signal having an appropriate amplitude. A feedback resistor 3 for setting the amplification factor (ie gain) of the transimpedance preamplifier 2, a postamplifier 4 for amplifying the output signal of the transimpedance preamplifier 2, a clock extraction circuit 5 for extracting a clock signal from the output signal of the postamplifier 4, The level shift circuit 6 (6a, 6b, 6c,...) That appropriately shifts the output signal level (amplitude) of the postamplifier 4 and the output clock signal of the clock extraction circuit 5 provided for each level shift circuit 6. Are appropriately delayed by clock delay circuits 7 (7a, 7b, 7c, ..) Provided for each level shift circuit 6 and clock delay circuit 7, and the signal level is “H” or “L” based on the output signals of these level shift circuit 6 and clock delay circuit 7. An identification circuit 8 (8a, 8b, 8c) for identifying the above, an arithmetic circuit 9 for counting and calculating these identification results and issuing an alarm signal based on the output of the identification results of each identification circuit 8, and an arithmetic circuit 9 is provided with an alarm output terminal 10 for outputting the issued alarm signal.
[0015]
Next, the operation of the abnormality detection circuit 20 will be described. For example, a digital light pulse signal whose intensity is randomly modulated at a constant pulse interval is input to the photoelectric conversion element 1 and converted into a current pulse signal as an electric signal. The current pulse signal is amplified to an appropriate amplitude by the transimpedance preamplifier 2 and the postamplifier 4.
[0016]
FIG. 2 shows an eye pattern of the current pulse signal after amplification. In this figure, the horizontal axis represents time, and the vertical axis represents amplitude (level). In the eye pattern, in order to identify whether the level (ie, amplitude) of the code in the signal is “H” (: large) or “L” (: small), ) To set the true identification point 11. For each code, the true discrimination point 11 and its signal level are compared. For example, if the signal level is equal to or higher than the discrimination threshold of the true discrimination point 11, the code is identified as “H” (: large). If the level is lower than the identification threshold value of the true identification point 11, the code is identified as “L” (: small).
[0017]
Identification of “H” or “L” at the true identification point 11 is performed by the main identification circuit 8a. In order to accurately perform this identification, the amplitude and time of the signal are offset before being input to the main identification circuit 8a. The amplitude offset is performed by the first level shift circuit 6a, and the time offset is performed by the first clock delay circuit 7a. That is, the first level shift circuit 6a adjusts the true identification point 11 to be approximately the center in the vertical direction (amplitude direction) of the predetermined eye pattern, and the first clock delay circuit 7a Similarly, the identification point 11 is adjusted so as to be approximately at the center in the horizontal direction (time axis direction) of the predetermined eye pattern. With these offsets, the identification threshold value and the identification timing in the main identification circuit 8a are set or adjusted. In the present embodiment, the first level shift circuit 6a is connected to the output of the postamplifier 4 and the input of the main identification circuit 8a, and the first clock delay circuit 7a is the output of the clock extraction circuit 5 and the main identification. Each is connected to the input of the circuit 8a.
[0018]
In the present invention, one or more reference identification points 12 are provided in addition to the true identification point 11. Identification of “H” or “L” at the reference identification point 12 is performed by reference identification circuits (8b, 8c,...). This identification is performed in parallel with the first level shift circuit 6a, the first clock delay circuit 7a, and the main identification circuit 8a, and with the same connection relationship, the level shift circuit (6b, 6c,...), The clock delay. Circuit (7b, 7c,...) And reference identification circuit (8b, 8c,...), Level shift amount of the level shift circuit (6b, 6c,...), And clock delay circuit ( 7b, 7c,..., And the clock delay amount of the first level shift circuit 6a (amplitude offset amount) and the clock delay amount of the first clock delay circuit 7a (time offset amount), respectively. This is realized by using different amounts. The reference identification point 12 (x mark in FIG. 2) can be shown as a point at a position different from the true identification point 11 (circle mark in FIG. 2) on the eye pattern shown in FIG. Each reference identification point 12 is set at a position separated from the true identification point 11 on the eye pattern in accordance with the level shift amount or the clock delay amount from the true identification point 11. Become. That is, each reference identification circuit (8b, 8c,...) Is binary-identified by a reference threshold or reference timing different from the identification threshold or identification timing (that is, large or small identification with respect to each threshold set). I do.
[0019]
The identification result obtained by each identification circuit 8 is input to the arithmetic circuit 9. The arithmetic circuit 9 has a counter function, and counts the number of cases where there is a difference between the identification result of the identification circuit of the true identification point 11 and the identification circuit of the reference identification point 12 within a predetermined period. The ratio of the count number when the identification results are different from the number of bits of all codes transmitted within the period is obtained. Since there is a correlation between the code identification result and the position of the identification point on the eye pattern, the ratio when the identification results at the true identification point 11 and the reference identification point 12 are different from each other, and the respective identification points on the eye pattern From the positional relationship, the code error rate at the true discrimination point 11 can be estimated. That is, it is possible to issue an alarm signal when the code error rate exceeds a predetermined value from the arithmetic circuit 9 by appropriately setting the positional relationship of the identification points, the count measurement period, and the count quantity of the difference between the identification results. It becomes.
[0020]
Embodiment 1 FIG.
As a first embodiment of the present invention, a case will be described in which one point whose level is different (for example, higher) than the true identification point 11 is used as the reference identification point 12. FIG. 3 shows an eye pattern in this case. As shown in the figure, here, as an example, the level of the true discrimination point 11 (ie, the discrimination threshold) is 2a lower than the level of the “H” signal, and the level of the reference discrimination point 12 (ie, the reference threshold) is set to “H”. The signal level is set a lower than the signal level. Hereinafter, the number of cases where the code error rate at the true discrimination point 11, the count measurement period, and the discrimination results at the true discrimination point 11 and the reference discrimination point 12 are different will be described. In the following, the signal transmission rate is 2.5 gigabits / second (Gb / s), the difference counting period (described later) by the arithmetic circuit 9 is 40 microseconds (μs), and the code error rate is 10 ^-3 A case will be described in which an alarm is issued when (0.001) is exceeded.
[0021]
In the eye pattern of FIG. 3, each signal is represented by a single line. However, considering the noise component included in the signal, the levels of the “H” signal and the “L” signal are probabilities based on normal distributions, respectively. It can be assumed that it spreads out. This is shown in FIG. In this figure, the vertical axis indicates the level, the horizontal axis indicates the probability, the upper curve indicates the probability distribution of the detection level of the “H” signal, and the lower curve indicates the probability distribution of the detection level of the “L” signal. . In this figure, the probability that the “L” signal is mistaken for the “H” signal (that is, the probability that the signal to be the “L” signal is identified as the “H” signal) is the integrated value of the probability distribution in the probability distribution range 13. (That is, the area of the region 13 in FIG. 4), and the probability that the “H” signal is mistaken for the “L” signal (that is, the probability that the signal to be the “H” signal is identified as the “L” signal) is a probability distribution. The integrated value of the probability distribution in the range 14 (that is, the area of the region 14 in FIG. 4). That is, in the normal distribution, the sum of the area of the probability distribution range 13 and the area of the probability distribution range 14 is the code error rate P (R) at the true identification point 11. Here, in the present embodiment, the true identification point 11 is set so that the areas of the probability distribution ranges 13 and 14 are substantially equal. That is, in other words, the probability that the “L” signal is mistaken as the “H” signal above the discrimination threshold (ie, the area of the region 13), and the probability that the “H” signal is mistaken as the “L” signal below the discrimination threshold. The level of the true discrimination point 11 (that is, the discrimination threshold) is set so that (that is, the area of the region 14) is substantially equal.
[0022]
The probability P (D) between the discrimination result at the reference discrimination point 12 and the discrimination result at the true discrimination point 11 is the sum of the probability distribution range 15 and the probability distribution range 13 as shown in FIG. Here, the probability of discriminating one of the binary signals (in this case, the “L” signal) from the other signal (in this case, the “H” signal) at the level of the reference identification point 12 (ie, the reference threshold) is very high. If this is assumed to be zero because it is low, the probability distribution range 13 and the probability distribution range 14 are substantially equal, and thus P (D) can be said to be the sum of the probability distribution range 15 and the probability distribution range 14. This corresponds to the upper probability of the normal distribution of the “H” signal at the reference identification point 12 (that is, the probability that the signal is an “H” signal at a level below the reference threshold in this case). Here, the code error rate P (R) is 10 ^-3 Consider the time of (0.001). In this case, the sum of the probability distribution range 13 and the probability distribution range 14 is 10 ^-3 Therefore, the probability distribution ranges 13 and 14 (that is, the upper probability in the normal distribution) are 5 × 10 5 respectively. ^-4 (0.0005). Upper probability is 5 × 10 in normal distribution ^-4 When the variance is σ, the level 2a of the true discrimination point 11 is 2σ = 3.3σ. At this time, the level a of the reference identification point 12 is half that of 1.65σ, and the upper probability of the normal distribution at this time is approximately 0.0495. Therefore, the code error rate is 10 ^-3 At this time, it can be said that the probability (difference occurrence probability) that a difference occurs between the identification result at the true identification point 11 and the identification result at the reference identification point 12 is 0.0495. The total number of codes transmitted during a predetermined period (for example, 40 microseconds) for operating the counter in the arithmetic circuit 9 is 100,000 bits when the transmission speed is 2.5 Gb / s. Therefore, the difference of the identification result is counted by the counter of the arithmetic circuit 9, and this is integrated for a predetermined period (40 microseconds), and the result is a number corresponding to the difference occurrence probability 0.0495 in this period, that is, 100,000 × If an alarm is issued with an abnormality when 0.0495 = 4950 count is exceeded, this is converted to a code error rate of 10 ^-3 An abnormality can be detected and an alarm can be issued when the value exceeds.
[0023]
Embodiment 2. FIG.
In this embodiment, as shown in FIG. 5, a point having a timing different from (for example, earlier) the true identification point 11 is set as a reference identification point 12, and the number of occurrences of a difference between large and small identification results with respect to the same identification threshold is counted. Issue an alarm. Hereinafter, this principle will be described.
[0024]
A signal that should be an identification threshold at the timing when the signal changes from “H” to “L” or “L” to “H” on both sides of the eye (ie, timings A and B in FIGS. 5 and 6). The probability distribution of the timing at which is actually the identification threshold is the same as that in the first embodiment. As in FIG. 6, the normal distribution is usually as shown in FIG.
[0025]
In this embodiment, the probability that the signal that should be the discrimination threshold at timing A is actually the discrimination threshold at a timing later than timing C (34 in FIG. 6), and the signal that should be the discrimination threshold at timing B is actually The timing C at which the probability of being an identification threshold at a timing earlier than the timing C (33 in FIG. 6) is equal is set as the identification timing. Further, the reference timing D of the reference identification point 12 is a timing that is different (for example, earlier) from the identification timing C, and has almost no probability that a signal that should become the identification threshold at the timing B actually becomes the identification threshold at that timing. (That is, the timing deviated from the probability distribution corresponding to B in FIG. 6). Note that the discrimination threshold values of the true discrimination point 11 and the reference discrimination point 12 are, for example, values just between the “H” level and the “L” level.
[0026]
In the situation where two identification points are set in this way, when a signal appears in accordance with the normal distribution shown in FIG. 6, the signal that should become the identification threshold at timing A is identified at a timing later than the reference timing D. The actual probability (the area of 36 in FIG. 6) serving as the threshold is the probability on the normal distribution that the signal becomes the identification threshold at the timing between the reference timing D and the identification timing C (the areas of 35 and 33 in FIG. 6). The total). However, if the code error rate (corresponding to the sum of the areas 33 and 34 in FIG. 6 on the normal distribution) increases, the signal becomes the discrimination threshold at the timing between the reference timing D and the discrimination timing C. The actual probability becomes greater than the probability (area 36 in FIG. 6) on the normal distribution that the signal that should be the discrimination threshold at timing A becomes the discrimination threshold at a timing later than the reference timing D.
[0027]
Here, as shown in FIG. 7, when the signal is at the “H” level at the reference timing D and at the “L” level at the identification timing C, and when the signal is at the “L” level at the reference timing D, If the signal is at the “H” level at the identification timing C, the signal becomes an identification threshold between the reference timing D and the identification timing C. That is, when the identification result for the identification threshold value at the reference identification point 11 is different from the identification result for the identification threshold value at the true identification point 12, the signal is identified at a timing between the identification timing C and the reference timing D. It becomes a threshold value.
[0028]
Therefore, among the number of all signals that change from “H” to “L” or “L” to “H” at the timing A, the ratio of the number of signals having different discrimination results at the two discrimination points 11 and 12 is: When the signal that should become the discrimination threshold at timing A becomes higher than the probability (area 36 in FIG. 6) on the normal distribution that becomes the discrimination threshold at a timing later than the reference timing D, the code error rate assumed in the normal distribution It can be identified that the error rate is higher.
[0029]
Here, the code error rate is 10 ^-3 Consider an example in which an alarm is issued when the threshold is exceeded. The code error rate on the normal distribution is 10 ^-3 , The probability that the signal that should become the identification threshold at timing A becomes the identification threshold at a timing later than timing C (34 in FIG. 6), and the signal that should become the identification threshold at timing B is identified at a timing earlier than timing C. Each of the threshold probabilities (33 in FIG. 6) is 5 × 10. ^-4 It becomes. Upper probability is 5 × 10 in normal distribution ^-4 Is when 3.3σ (σ is a dispersion value). When the reference timing D (b) is set for the identification timing C (2b) as shown in FIG. 5, the probability above the reference timing D (b = 1.65σ) (the area of 36 in FIG. 6) is 0. 0495. Of all the signal transition patterns before and after timing A (“H” → “H”, “H” → “L”, “L” → “H”, “L” → “L”), “H” Since the probability that the value changes from “L” or “L” to “H” is ½ of the probability, when the signal follows a normal distribution, two identification points 11 and 12 out of all signals. The probability when the identification results are different (the total area of 33 and 34 in FIG. 6) is 0.02475 (= 0.0495 / 2). This corresponds to ½ of the case of the first embodiment. The total number of codes transmitted during a predetermined period (for example, 40 microseconds) for operating the counter by the counter of the arithmetic circuit 9 is 100,000 bits when the transmission speed is 2.5 Gb / s. Therefore, the difference of the identification results is counted by the counter of the arithmetic circuit 9, and this is integrated for a predetermined period (for example, 40 microseconds), and the result is a number corresponding to the difference occurrence probability 0.02475 in this period, that is, 100,000. × 0.02475 = 2 When the count exceeds 2475, the code error rate is 10 ^-3 An alarm may be issued as an abnormality exceeding. In this embodiment, the true identification point 11 is identified by the main identification circuit 8a, and the reference identification point 12 is identified by the reference identification circuit 8b.
[0030]
Embodiment 3 FIG.
As shown in FIG. 8, as the third embodiment of the present invention, reference is made to both a point with a different (for example, higher) level and a point with a different (for example, earlier) timing with respect to the true identification point 11. By using the identification point 12 and counting the number of occurrences of the difference between the reference identification point 12 and the true identification point 11 in the same manner as in the first and second embodiments, a predetermined code error rate was exceeded. An alarm can be issued in case. In this case, the first embodiment and the second embodiment are combined, the difference in the identification result is counted for each of the level direction and the time axis direction, and both counters exceed a predetermined number of times, or either By issuing an alarm when the number of counters exceeds a predetermined number, it is possible to realize an alarm when a predetermined code error rate is exceeded.
[0031]
【The invention's effect】
As described above, according to the present invention, it is possible to realize a circuit that detects an optical signal abnormality according to a code error rate based on an identification result for an identification threshold and an identification result for a reference threshold different from the identification threshold.
[0032]
Further, according to the present invention, the number of times when the magnitude of the electrical signal intensity is different from the magnitude of the discrimination threshold and the magnitude of the discrimination result relative to the reference threshold is accumulated for a predetermined period. By making an abnormality when the value is exceeded, a circuit that detects an abnormality of the optical signal according to the code error rate can be realized.
[0033]
Further, according to the present invention, when the discrimination threshold is set in the main discrimination circuit, the probability that a signal that should be smaller than the discrimination threshold is identified as a signal that is larger than the discrimination threshold, and greater than the discrimination threshold. By setting so that the probability that the signal to be identified is identified as a signal smaller than the identification threshold is equal, it is possible to realize a circuit that detects an optical signal abnormality according to the code error rate.
[0034]
Further, according to the present invention, it is possible to realize a circuit that detects an abnormality of an optical signal according to a code error rate based on an identification result at an identification timing and an identification result at a reference timing different from the identification timing.
[0035]
Further, according to the present invention, the number of times when the magnitude discrimination result with respect to the discrimination threshold at the discrimination timing of the intensity of the electric signal is different from the magnitude discrimination result with respect to the discrimination threshold at the reference timing is further set to a predetermined number. It is possible to realize a circuit that detects an abnormality of the optical signal according to the code error rate by integrating the periods and setting an abnormality when the number of times of accumulation exceeds a predetermined value.
[0036]
Further, according to the present invention, when the identification timing is set in the main identification circuit, the electrical signal that should become the identification threshold at the early signal change timing in the eye is actually higher than the identification timing. The probability that the discrimination threshold is set at a later timing and the probability that the electrical signal that should be the discrimination threshold at the late signal change timing in the eye actually becomes the discrimination threshold at an earlier timing than the discrimination timing are set to be equal. By doing so, it is possible to realize a circuit that detects an abnormality of the optical signal according to the code error rate.
[0037]
Further, according to the present invention, based on the identification result for the identification threshold at the identification timing, the identification result for the reference threshold different from the identification threshold, and the identification result at the reference timing different from the identification timing, A circuit for detecting a signal abnormality can be realized.
[0038]
Further, according to the present invention, an optical receiver including the above-described abnormality detection circuit is realized.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an anomaly detection circuit for an optical receiver according to an embodiment of the present invention.
FIG. 2 is a diagram showing an eye pattern of a signal and an identification point in the optical receiver abnormality detection circuit according to the embodiment of the present invention.
FIG. 3 is a diagram showing an eye pattern and an identification point of an input signal in the optical receiver abnormality detection circuit according to the first embodiment of the present invention.
FIG. 4 is an explanatory diagram showing signal level probability distribution and identification point setting in the optical receiver anomaly detection circuit according to the embodiment of the present invention;
FIG. 5 is a diagram showing an eye pattern of an input signal and an identification point in the optical receiver abnormality detection circuit according to the second embodiment of the present invention.
FIG. 6 is a diagram showing a probability distribution of timing when a signal value becomes an identification threshold value in the optical receiver abnormality detection circuit according to the second embodiment of the present invention.
FIG. 7 is an explanatory diagram showing a signal serving as an identification threshold at a timing between an identification timing and a reference timing in the optical receiver abnormality detection circuit according to the second embodiment of the present invention.
FIG. 8 is a diagram showing an eye pattern of an input signal and an identification point in an optical receiver abnormality detection circuit according to a third embodiment of the present invention.
FIG. 9 is a block diagram showing a conventional optical receiver abnormality detection circuit.
[Explanation of symbols]
1 photoelectric conversion element 2 transimpedance preamplifier 3 feedback resistor 4 post amplifier 5 clock extraction circuit 6 level shift circuit 7 clock delay circuit 8 identification circuit 8a main identification circuit 8b, 8c, etc. Identification circuit, 9 arithmetic circuit, 11 true identification point, 12 reference identification point, 20 optical receiver abnormality detection circuit.

Claims

A main identification circuit that compares the intensity of the electrical signal photoelectrically converted from the optical signal by the photoelectric conversion element with a predetermined identification threshold;
A reference identification circuit that compares the intensity of the electrical signal with a reference threshold different from the predetermined identification threshold;
An arithmetic circuit that detects a signal abnormality based on a comparison result between the main identification circuit and the reference identification circuit;
Equipped with a,
The arithmetic circuit integrates the number of times when the magnitude discrimination result of the electric signal strength with respect to the discrimination threshold is different from the magnitude discrimination result with respect to the reference threshold for a predetermined period, and the accumulated number of times is the given code error. The abnormality detection threshold calculated from the probability that a difference occurs in each comparison result of the main identification circuit and the reference identification circuit determined according to the rate and the total number of codes transmitted during the predetermined period An optical receiver abnormality detection circuit for detecting a case as a signal abnormality .

The discrimination threshold includes a probability that a signal that should be smaller than the discrimination threshold is identified as a signal larger than the discrimination threshold when the discrimination threshold is set in the main discrimination circuit, and a signal that should be larger than the discrimination threshold is smaller than the discrimination threshold. 2. The anomaly detection circuit for an optical receiver according to claim 1, wherein the anomaly detection circuit is set so that the probability of being identified as a signal is equal.

A main identification circuit that compares the intensity of the electrical signal photoelectrically converted from the optical signal by the photoelectric conversion element at a predetermined identification timing with a predetermined identification threshold;
A reference identification circuit that compares the strength of the electrical signal with a predetermined identification threshold at a reference timing different from the predetermined identification timing;
An arithmetic circuit that detects an abnormality of the signal based on the identification results of the main identification circuit and the reference identification circuit;
Equipped with a,
The arithmetic circuit accumulates the number of times when the magnitude discrimination result with respect to the discrimination threshold at the discrimination timing of the intensity of the electric signal is different from the magnitude discrimination result with respect to the discrimination threshold at the reference timing for a predetermined period, and accumulates the number of times. The number of times is calculated from the probability that a difference occurs in each comparison result of the main identification circuit and the reference identification circuit determined according to a given code error rate and the total number of codes transmitted during the predetermined period. An optical receiver abnormality detection circuit for detecting a signal abnormality when the abnormality detection threshold is exceeded .

When the identification timing is set in the main identification circuit, the electrical signal that should become the identification threshold at the earlier signal change timing in the eye actually becomes the identification threshold at a timing later than the identification timing. The probability is set to be equal to the probability that the electrical signal that should become the discrimination threshold at the later signal change timing in the eye is actually the discrimination threshold at a timing earlier than the discrimination timing. An abnormality detection circuit for an optical receiver according to claim 3 .

A main identification circuit that compares the intensity of the electrical signal photoelectrically converted from the optical signal by the photoelectric conversion element at a predetermined identification timing with a predetermined identification threshold;
A first reference identification circuit that compares the intensity of the electrical signal with a reference threshold different from the predetermined identification threshold at a predetermined identification timing;
A second reference identification circuit that compares the strength of the electrical signal with the predetermined identification threshold at a reference timing different from the predetermined identification timing;
An arithmetic circuit that detects an abnormality of the signal based on the identification results of the main identification circuit, the first reference identification circuit, and the second reference identification circuit;
Equipped with a,
The arithmetic circuit integrates the number of times when the magnitude discrimination result of the electric signal strength with respect to the discrimination threshold is different from the magnitude discrimination result with respect to the reference threshold for a predetermined period, and the accumulated number of times is the given code error. A first abnormality detection threshold value calculated from a probability that a difference occurs in each comparison result of the main identification circuit and the reference identification circuit determined according to a rate and the total number of codes transmitted during the predetermined period Or the number of times when the discrimination result of the magnitude of the electric signal strength with respect to the discrimination threshold at the discrimination timing differs from the discrimination result with respect to the discrimination threshold at the reference timing is accumulated for a predetermined period, and the integration is performed. The probability of occurrence of a difference between the comparison results of the main identification circuit and the reference identification circuit determined according to a given code error rate and Optical receiver dexterity abnormality detection circuit for detecting a case where at least one is generated as a signal of abnormal when exceeding the second abnormality detection threshold value is calculated from the total number of codes to be transmitted during the period.

An optical receiver including an optical receiver dexterity abnormality detection circuit according to any one of claims 1 to 5.