JP3913303B2 - Magnetic disk storage device - Google Patents

Description

上記のタップ重みを有する９タップウィンドウ式ＦＩＲフィルタの出力の結果、図１８に示すようにリプルが除去される。さらに図１９に示すように、ウィンドウ式９タップＦＩＲフィルタはその完璧な線形位相応答を保持する。ただし、プログラム可能ＦＩＲフィルタ８８のタップ重みにハミング・ウィンドウなどのウィンドウ機能を適用することは、非ゼロのＤＣ利得とある程度の低周波応答の増加を見込んでいることに留意されたい。
【００５３】
次に図２０ないし図３０に移行すると、ＭＲエレメント７２を使用して磁気データ記憶ディスク２４から獲得した読戻し信号の高域通過フィルタリングを使用するデータ記憶システムに特に適した信号分離／復元モジュール７６のもう１つの実施例が示されている。この実施例は、新たに設計したデータ記憶システムならびに改装したシステムで使用することができる。図１に示すＡＥモジュール７４などのアナログＡＥモジュール７４の設計では、磁気信号成分の周波数範囲以下の読戻し信号の信号内容を排除するために前置増幅器とともに高域通過フィルタリングを含むことが望ましい場合が多い。ＡＥモジュール７４の高域通過フィルタリング挙動により、複合読戻し信号の熱信号成分が振幅と位相の両面でひずむ。この熱信号ひずみは、使用する特定のＡＥモジュールの周波数および位相応答に応じて重大度が変動する。
【００５４】
一例として、ＡＥモジュール７４での使用に適した高域通過フィルタは、約５００ＫＨｚの遮断周波数を有し、非線形位相挙動を示すことができる。しかし、意味のある熱信号情報に関連する周波数は、通常、２００ＫＨｚ以下であり、１０ＫＨｚから約１００ＫＨｚの範囲である。約５００ＫＨｚの遮断周波数を有する高域通過フィルタは、読戻し信号の熱信号成分の振幅および位相をかなりひずませることに留意されたい。しかし、読戻し信号の磁気信号成分は、引き続き高域通過フィルタの影響を受けない。というのは、一般に磁気信号の周波数範囲が高域フィルタ遮断周波数の２０〜４０倍程度であるからである。
【００５５】
図２０および図２１には、高域通過フィルタリング挙動を示す典型的なアナログＡＥモジュール７４の絶対値および位相応答を示すグラフがそれぞれ示されている。この高域通過フィルタは、約５００ＫＨｚの遮断周波数を有する。図２０および図２１に示す絶対値および位相応答と５００ＫＨｚの単一極を有するＡＥモジュール７４の有効高域通過フィルタのアナログ伝達関数と同等のディジタル版は、次のように定義することができる。
【数１】

■F
ｂ_h（１）＝．９８７６
ｂ_h（２）＝−．９８７６
ａ_h（２）＝−．９７５２
【００５６】
ＡＥモジュール７４の高域通過フィルタリング挙動によってもたらされた熱信号の絶対値および位相のひずみは、高域通過フィルタのそれとは逆の伝達関数を有する逆フィルタの使用によって効果的に除去される。ＡＥモジュール７４から出力された読戻し信号を逆フィルタに通過させると、熱信号が振幅および位相の両面でその元の形式に復元される。たとえば、式［１］の上記の伝達関数を有する高域通過フィルタを通過した読戻し信号の条件調整を行うための逆フィルタの伝達関数は次式で表される。
【数２】

Ｏ０５７】
ＡＥモジュール７４の有効高域通過フィルタと式［２］で前述した逆フィルタの絶対値および位相応答について、図２２および図２３にそれぞれ示す。特に、逆フィルタとＡＥモジュール７４の有効高域通過フィルタの絶対値応答は、図２２の曲線１７０および１７２としてそれぞれ示す。また、逆フィルタと有効高域通過フィルタの位相応答は、図２３の曲線１７６および１７４としてそれぞれ示す。
【００５８】
一実施例では、高域通過フィルタリング読戻し信号の熱信号内容を復元するために信号分離／復元モジュール７６内の逆フィルタとして、無限インパルス応答（ＩＩＲ）フィルタが使用される。ＩＩＲフィルタのインパルス応答は、インパルス応答が有限持続期間を有するＦＩＲフィルタとは対照的に、持続時間が無限である。また、完璧な線形位相応答を示す可能性のあるＦＩＲフィルタとは異なり、ＩＩＲフィルタの位相応答は、特に帯域のエッジでは非線形になる。代替実施例ではアナログ・フィルタを使用することができるが、ＩＩＲフィルタは、アナログＡＥモジュール７４の高域通過フィルタリング挙動によってひずんだ熱信号の振幅および位相を復元するために、逆フィルタとしての使用に適したいくつかの利点をもたらす。
【００５９】
図２４に示す信号流れ図は、逆フィルタとして構成された１次ＩＩＲフィルタを表している。上記の式［２］で示される伝達関数を有する１次ＩＩＲ逆フィルタ用の図２４の信号流れ図に関連する係数は、以下の通りである。
ａ₁＝．９８７６
ａ₂＝−．９８７６
ｂ₁＝．１
ｂ₂＝−．９７５２
信号分離／復元モジュール７６での逆フィルタとしての使用に適したＩＩＲフィルタの設計、実現、プログラミングの詳細については、E. C. Ifeachor、B. W. Jervisによる"Digital Signal Processing"（Addison-Wesley Publishing Company, Inc. 1993）を参照されたい。
【００６０】
図２５ないし図２７には、高域通過フィルタを通過した読戻し信号の熱信号成分の元の振幅および位相を復元するための逆フィルタの有効性を実証するために使用する３通りの波形が示されている。図２５には、ＭＲヘッドを使用してデータ記憶ディスク表面のピットを走査することによって検出した読戻し信号を示す。図２５に示す磁気読戻し信号は、２０ＭＨｚの書込み周波数で書き込まれたトラックから検出されたものである。この磁気読戻し信号は８ビットの解像度で１００ＭＨｚでサンプリングされている。図２６に示す信号は、図２５の読戻し信号の計算ピーク間絶対値を表している。さらに、図２６に示す信号は、ピットの上をＭＲ読取りエレメントが通過することにより振幅の大幅低減を明確に実証するような、図２５に示す読戻し信号の磁気信号成分を表している。図２７は、ＡＥモジュール７４の有効高域通過フィルタを通過した後の図２５の読戻し信号の熱信号成分を示している。図２６および図２７の波形を比較すると、読戻し信号の磁気信号成分と熱信号成分が互いに密接に対応しないことが分かる。このように熱信号と磁気信号があまり対応していないことは、アナログＡＥモジュール７４の有効高域通過フィルタリング特性による熱信号のひずみの結果であり、それによって熱信号が効果的に区別されている。
【００６１】
信号分離／復元モジュール７６の逆フィルタは、図２８に示すように熱信号１６２の振幅と位相を復元する。ただし、図２８に示す熱信号と磁気信号は、以下に詳述するようにヘッド／ディスク間隔信号として示されていることに留意されたい。磁気信号と復元熱信号は、高域通過フィルタリング熱信号が逆フィルタを通過した後、互いに密接な対応を示すことが分かる。
【００６２】
図８に関連して前述したように、ＭＲヘッドで誘導された熱信号は、ヘッド／ディスク間隔の関数として変化する。したがって、熱信号に含まれる情報を使用すると、ディスクの表面凹凸の変動を検出することができる。熱信号を使用して、ピット、打こん、こぶ、熱によるでこぼこ、汚染粒子など、様々な表面フィーチャを検出することができる。このような表面フィーチャは、熱信号を使用して様々なタイプの情報を導出するために意図的にディスク表面に取り入れることができることに留意されたい。
【００６３】
たとえば、熱信号を使用してトラックおよびセクタ位置を判定するために、同心かつ半径方向に細長いくぼみをディスク表面に含めることができる。また、熱信号を使用して、ディスク表面の詳細な表面凹凸マッピングも実施することができる。読戻し信号から抽出された熱信号の可用性は、広範囲の応用分野で使用するために有利に活用できることに留意されたい。他の例としては、熱信号を使用してヘッド／ディスク間隔の測定値を求めるために、ＭＲヘッドの熱応答を較正するために既知の深さまで１つまたは複数のくぼみをディスク表面に形成することもできる。
【００６４】
ディスクの表面と変換器との間の間隔の変化を決定するために、読取り／書込み変換器が発生した磁気読取り信号を使用することは、当業者には既知のことである。磁気読戻し信号を使用してヘッド／ディスク間隔を判定するためのこのような方法の１つは、調波比浮動高さ（ＨＲＦ）クリアランス・テストと呼ばれる。ＨＲＦテストは、磁気ヘッド／ディスク間隔信号を使用して現場またはデータ記憶システム・ハウジング内で実行される、変換器を支持するスライダの浮動高度を測定するための既知の方法である。ＨＲＦ方法については、本発明の出願人に譲渡され、参照により本明細書に組み込まれる、米国特許第４７７７５４４号に記載されている。このＨＲＦ測定方法は、読戻し信号のスペクトル内の２本のスペクトル線の割合を連続して瞬時に測定する方法である。両方の瞬間スペクトル線の振幅は、ＭＲ変換器のすぐ下にある記録媒体の同一ボリューム・エレメントに関係する。また、このＨＲＦ測定方法は、磁気読戻し信号を使用してディスク表面に対する瞬間ヘッド・クリアランスを決定することができる。
【００６５】
一実施例によれば、ＭＲヘッドで誘導された読戻し信号の熱信号成分は、ヘッド／ディスク間隔の変化を定性的に決定するために使用される。他の実施例では、ヘッド／ディスク間隔の定量的決定に対応するように、磁気信号を使用して熱信号が較正される。図２９を参照すると、同図には、読戻し信号を処理して磁気および熱それぞれのヘッド／ディスク間隔情報を得るためのシステムがブロック図形式で示されている。読戻し信号は、ＭＲエレメント７２によってディスク表面２４から検出される。例示のため、読戻し信号は磁気信号成分と熱信号成分の両方を含む複合信号であると想定し、磁気信号成分が欠けている読戻し信号はヘッド／ディスク間クリアランスを決定するために有用な熱信号成分を含むものと理解されている。ＭＲエレメント７２によって検出された読戻し信号は、ＡＥモジュール７４に、次に高域通過フィルタ１５０に通信される。高域通過フィルタ１５０は、ＡＥモジュール７４外部の構成要素として示されているが、全体的にＡＥモジュール７４の高域通過フィルタリング挙動を表すように設けられている。有効高域通過フィルタ１５０の伝達関数はＨ₀として示されている。高域通過フィルタリング読戻し信号のディジタル化サンプルを生成するために、高域通過フィルタ１５０からの出力信号がアナログ・ディジタル変換器１５１によってサンプリングされる。
【００６６】
図２９に示すように、熱信号抽出フィルタ１５７の出力側の点１５９に示す熱信号は、前述の方法のいずれを使用しても発生することができる。たとえば、ディジタル化読戻し信号は、ＡＥモジュール７４の高域通過フィルタ１５０によってもたらされるひずみを矯正する逆フィルタ１５６に通信することができる。逆フィルタ１５６の伝達関数はＨ₀ ^-1として示されている。次に、熱信号抽出フィルタ１５７によって熱信号が抽出されるが、このフィルタはＦＩＲフィルタにすることができる。逆フィルタ１５６と熱信号抽出フィルタ１５７は、高域通過フィルタ１５０によってひずんだ熱信号を復元するために単一のＩＩＲフィルタに実現できることに留意されたい。あるいは、高域通過フィルタ１５０より前のある点で読戻し信号をタッピングし、熱信号抽出フィルタ１５７に入力することができるが、このフィルタは上記で詳述したようにＦＩＲフィルタにすることができる。熱信号抽出フィルタ１５７によって抽出された熱信号は平均フィルタ１５８に通信され、次にこのフィルタはヘッド／ディスク間隔に直線的に関連する熱間隔信号１６２を発生する。平均フィルタ１５８は、ディジタル動作平滑平均化フィルタである。
【００６７】
アナログ・ディジタル変換器１５１の出力側に出力される読戻し信号は、読戻し信号のピーク間振幅を検出し、読戻し信号から磁気信号成分を抽出する、ＦＩＲフィルタなどの振幅検出器１５２にも通信することができる。磁気信号の対数は、ログ装置１５４で磁気信号を通過させることによって得られるが、この装置はヘッド／ディスク間隔に直線的に関連する磁気信号を発生する。磁気間隔信号１６０と熱間隔信号１６２の両方をそれぞれ抽出すると、熱信号を較正することができる。というのは、磁気較正は既知であり、その信号の記録済み波長のみに依存しているからである。磁気間隔信号１６０の負数（または逆数）と抽出した熱信号１６２がヘッド／ディスク間隔（ｙ）に直線的に比例することに留意することは重要である。
【００６８】
図２８には、熱信号抽出フィルタ１５７と平均フィルタ１５８によって処理された熱間隔信号１６２が、振幅検出器１５２とログ装置１５４によって処理された磁気間隔信号１６０とともに示されている。ただし、線形化磁気間隔信号１６０は通常、ピーク間信号の対数を取ることによって計算され、周知のウォーレス式により磁気間隔変化に対する出力電圧変化の既知の感度が掛けられることに留意されたい。また、図２８では、熱間隔信号１６２に関連する信号の高さの差とそれよりわずかに長い時間定数を除き、磁気間隔信号１６０と熱間隔信号１６２がディスク表面のピットを表すことが分かる。以下に詳述するように、真のヘッド／ディスク間隔を正確に反映するため、線形化磁気間隔信号１６０を使用して熱間隔信号１６２を較正することができる。
【００６９】
本発明の重要な利点は、ＭＲエレメント７２の熱応答を使用して、現場またはデータ記憶システムのハウジング内でヘッド／ディスク間隔の変化を検出できる能力に関する。ＭＲエレメント７２の熱応答を使用する現場でのヘッド／ディスク間隔測定は、ディスク製造テストおよびスクリーニングのため、ならびに、現場でデータ記憶システムの耐用年数中に予測障害分析（ＰＦＡ）を実行するために有用である。さらに、熱間隔信号１６２を使用すると、データ記憶ディスクの表面とのヘッドの接触を検出することができる。
【００７０】
次に図３０を参照すると、ＭＲヘッドと局部的な熱によるでこぼこ（ＴＡ）との接触など、ヘッド／ディスク接触事象の場合の磁気間隔信号１６０と熱間隔信号１６２の両方が示されている。磁気間隔信号１６０は、磁気信号の対数を取ることによって線形化されている。また、熱間隔信号１６２は、前述の逆フィルタリング手法を使用することによって決定されている。ディスクのでこぼこによってＭＲエレメント７２がディスクの表面から上方向に変位すると、ＭＲエレメント／ディスク間隔が増加することが分かる。磁気間隔信号１６０と熱間隔信号１６２の両方は、０マイクロ秒から約２５マイクロ秒の間でヘッド／ディスク間隔がこのように徐々に増加すること示している。ＭＲエレメント７２がでこぼこの上を通過した後、ＭＲエレメント７２がその定常状態の浮動高度に戻る前にある程度の量の空気軸受け（ヘッド／ディスク間隔）変調が発生する。この空気軸受け変調は、約３５マイクロ秒から始まり、７０マイクロ秒まで継続することが図３０で分かる。
【００７１】
磁気間隔信号１６０と熱間隔信号１６２の波形の特徴の類似点は、テスト・ベンチ機器または外部テスタに頼らずに現場でヘッド／ディスク接触を検出するために熱間隔信号１６２を使用できることを実証している。ただし、特定のデータ記憶システムに必要な逆フィルタの特性がＡＥモジュール７４の高域通過フィルタ１５０の極位置に依存することに留意されたい。ＩＩＲフィルタまたはＦＩＲフィルタを使用する実施例の場合、係数またはタップ重みのみを変更する必要がある。ＩＩＲフィルタを使用する実施例では、この変更は、高域通過フィルタ１５０の極周波数で変動が発生した場合に適応的または動的に行うことができる。通常、このような変動は、温度などの変化につれて発生する。ただし、ここに記載する逆フィルタ１５０は１次ＩＩＲ構造に限定されてないことに留意されたい。
【００７２】
一般に、ＭＲヘッドで誘導された熱信号の絶対値は、ＭＲヘッドで使用する特定のＭＲエレメントの関数である。たとえば、使用する製造プロセスおよび材料の変動は、ＭＲエレメントの応答の変動の原因となる。したがって、ＭＲヘッドの熱応答を使用してヘッド／ディスク間隔の変化を正確に判定するために、磁気応答を使用して現場で熱応答を較正することが望ましい。たとえば、正確な磁気間隔情報は、周知のウォーレス間隔損失式を使用して求めることができる。半径方向のトレンチまたはピットなどの較正くぼみをランディング・ゾーンに形成して使用すると、現場での熱間隔較正を実行するために熱信号および磁気信号両方の変調を発生することができる。そのトレンチについて、磁気間隔を正確に決定することができ、ＭＲエレメントの熱電圧応答を較正するために使用することができる。
【００７３】
もう１つの方法は、ＨＲＦ方法またはその他の同様の方法を使用して得られた磁気ヘッド／ディスク間隔測定値と、熱クリアランス測定値とを組み合わせることを含む。この結合テストによれば、熱および磁気の同時「スピンダウン」が行われ、それにより、２つのディスク速度間の熱電圧変化が２つのディスク速度間の既知の（ＨＲＦ）間隔変化と比較される。しかし、ディスク回転速度での熱信号の回復は、高域通過フィルタリング挙動を有するＡＥモジュール７４を使用するシステムでは達成するのが困難な場合がある。というのは、高域通過遮断周波数は、通常、ディスク回転周波数より数桁も大きいからである。
【００７４】
ＭＲヘッドの熱応答のヘッド／ディスク間隔較正は、ＡＥモジュール７４の高域通過周波フィルタリング特性により、より複雑になっている。高域通過フィルタリング挙動を有するＡＥモジュール７４の伝達関数Ｈ_AE（ｓ）は、一般に以下のように１次近似値まで表すことができる。
【数３】

式中、Ｋ_AEは記録周波数でのＡＥモジュール７４の利得であり、「ａ」はＡＥモジュール７４に取り入れられた有効高域通過フィルタ用の遮断周波数である。利得の典型的な値はＫ_AE＝１７０であるが、ＡＥモジュール７４の利得は一般に変動が大きい。典型的な遮断周波数「ａ」は約３２５ＫＨｚであり、一般に＋／−１２５ＫＨｚという大きい許容誤差がそれに関連している。
【００７５】
表面分析スクリーニング中に検出されるこぶ状の表面欠陥の対象周波数は、通常、１０ＫＨｚから１００ＫＨｚの範囲である。ＭＲヘッドの熱応答はこのような周波数を直接変換するが、磁気応答は磁気記録搬送周波数のために２０ＭＨｚの範囲にこれらの周波数をシフトする。ＡＥモジュール７４の高域通過特性は、４００ＫＨｚ以下の周波数について熱応答内のすべてのこぶ外乱振幅を様々な量だけ減衰するが、磁気応答は影響を受けない。熱応答の減衰を復元するため、何らかの形式の集積化を適用する必要がある。この復元プロセスは、ＡＥモジュール７４の伝達関数Ｈ_AE（ｓ）とは逆の伝達関数Ｈ_INV（ｓ）を有する逆フィルタを使用することによって実施することができる（すなわち、Ｈ_INV（ｓ）＝１／Ｈ_AE（ｓ））。たとえば、７２００ＲＰＭで回転するディスクからデータを読み取るＭＲヘッドからの最低周波数は１２０Ｈｚであり、ディスク表面のこぶを検出するための最低周波数は約１０ＫＨｚである。したがって、４００ＫＨｚではゼロ（「ａ」）で５ＫＨｚでは極（「ｂ」）になる進み／遅れフィルタなどの擬似逆フィルタの方がこの応用分野にはより適している可能性がある。
【００７６】
擬似逆フィルタは、以下の形式の伝達関数を備えているものと思われる。
【数４】

また、ＡＥモジュール７４とカスケードされた擬似逆フィルタの全体的な伝達関数は以下のようになる。

したがって、上記の補正された伝達関数Ｈ（ｓ）は遮断周波数が５ＫＨｚの高域通過フィルタになり、これは、ひずみのない形式でディスク表面のこぶに関連する周波数を通過するのに妥当なものである。
【００７７】
高域通過遮断周波数「ａ」の変動が大きいので、復元した熱応答で大きな変動が発生する可能性がある。各ＭＲヘッドの高域通過遮断周波数「ａ」、利得Ｋ_AE、感度［ｎｍ／ｍｖ］を正確に推定することは、信頼性の高い較正にとって重要なことである。ＡＥモジュール７４の低周波（すなわち、１２０Ｈｚ以下）応答が欠落しているため、この熱較正プロセスは、参照の基準として別の方法によってサポートされる場合もある。このサポート方法は、ここでは磁気読戻し信号変調（ＲＳＭ）方法と呼ぶが、ウォーレス間隔損失手法に基づいてヘッド／ディスク磁気間隔の変化を決定するための既知の自己較正方法である。効果的な熱較正手順は、部分的にＲＳＭ方法に基づいており、較正くぼみを含むトラックの読戻し信号の熱成分と磁気成分の両方を検査するためにアクチュエータがランディング・ゾーン内のクラッシュ止めに押し付けられている間に最初に実行される。
【００７８】
気データと熱データの両方が得られるように研磨およびスパッタリング・プロセスが施される。ただし、ランディング・ゾーンに形成されたくぼみは、人造こぶ（manufactured bump）として形成することもできることに留意されたい。しかし、こぶは、ヘッド／ディスク干渉（ＨＤＩ）を発生する可能性がより高く、ヘッド／ディスク・クラッシュの結果としてヘッドまたはディスクに永続的な損傷を加える可能性がある。また、こぶは、ヘッドの上昇や空気軸受け変調を発生する可能性があるので、永続較正場所としての使用には不適当だと思われる。対照的に、「純粋な」ピットは、ヘッドの上昇や空気軸受け変調を引き起こさない。また、ディスク基板表面上の較正トレンチは、製造費用も安価である。
【００７９】
構成手順の一実施例について述べる前に、較正プロセスに関連するいくつかの変数を定義すると有用であると思われる。ただし、ＬＦ（低周波）という用語は、回転速度が７２００ＲＰＭのときの１２０Ｈｚなど、ディスク回転周波数（ＲＰＭ／６０）程度の周波数を指すことに留意されたい。式６および式７を参照すると、Ｖ_TH（ＬＦ）という用語は、較正ピットから得られたデータを除き、ランディング・ゾーンのＡＥモジュール復元熱電圧（基線）応答の回転ごとの平均を表し、通常、ミリボルト（ｍｖ）で表される。Ｖ_TH（Ｐｉｔ）という用語は、複数の回転について取られたランディング・ゾーンの較正ピットから発生したＡＥモジュール復元平均熱電圧ピークであり、通常、ミリボルト（ｍｖ）で表される。δ_HRF（ＬＦ）という用語は、較正ピットから得られたデータを除き、ランディング・ゾーンのＲＳＭヘッド／ディスク間隔の回転ごとの平均推定値を表し、通常、ナノメートル（ｎｍ）で表される。最後に、δ_HRF（Ｐｉｔ）という用語は、複数の回転について取られたランディング・ゾーンの較正ピットから発生した平均ピークＨＲＦヘッド／ディスク間隔であり、通常、ナノメートル（ｎｍ）で表される。
【００８０】
提案している熱較正プロセスは、ヘッド／ディスク間隔の回転ごとの平均低周波（ＬＦ）ＲＳＭ推定値δ_HRF（ＬＦ）と、これに対応する平均熱基線電圧Ｖ_TH（ＬＦ）の使用に基づくものであるが、これらはアクチュエータがランディング・ゾーンのクラッシュ止めにもたれた状態で共同で得られ、較正ピットから得られるデータは除く。除外した熱および磁気ピット・データの平均ピーク値により、Ｖ_TH（Ｐｉｔ）とδ_HRF（Ｐｉｔ）が得られる。
【００８１】

フヘッドの「ＡＣ」較正係数Ｃ（ｉ）は、次式から求めることができる。
【数６】
ピットの場合、δ_HRF（ＬＦ）＞δ_HRF（Ｐｉｔ）とδ_TH（ＬＦ）＞δ_TH（Ｐｉｔ）という条件が当てはまることが分かるはずである。ｉ番目の熱ヘッド／ディスク間隔の近似式は次のようになる。
δ_TH（ｉ）＝δ_HRF（ＬＦ）＋Ｃ（ｉ）・ΔＶ_TH ［７］
式中、ΔＶ_TH＝Ｖ_TH（ｄｅｆｅｃｔ）−Ｖ_TH（ＬＦ）である。こぶの場合は、ヘッド／ＭＲエレメントの接触がないと想定し、冷却が存在するので、ΔＶ_TH＜０になる。ピットの場合は、ヘッド／ディスク分離が増加するのでＭＲエレメントの加熱が発生し、したがって、ΔＶ_TH＞０になる。熱ヘッド／ディスク間隔近似式は較正ピットの位置（すなわち、この場合は内径ランディング・ゾーンであるが、ロード／アンロード・ディスク・ドライブの場合は外径になるはずである）で較正されているので、欠陥が発生しているトラック半径で平均ＲＳＭヘッド／ディスク間隔δ_HRF（ＬＦ）を更新することによって、正確さの改善を実現することができる。製造スクリーニング中にこれを実施するため、欠陥半径に磁気トラックが書き込まれるはずである。
【００８２】
ＭＲヘッドの熱応答の較正がなくても、読戻し信号の熱信号成分を使用すると、ディスク表面特性の定量分析ではなく定性分析を行うことができる。したがって、固有の正規化手法を使用してディスク欠陥を検出するために、ディスク表面分析スクリーニングを行うことができる。このような手法の１つは、そこからクリップ・レベル（すなわち、障害しきい値）が導出される基準としてディスク・トラック上の固有の「背景」熱信号情報を使用することに基づくものである。ディスク表面に磁気コーティングを施さなくても、ディスクの表面凹凸の定量評価と定性評価の両方を実行できることに留意することは重要である。したがって、磁気コーティングのないディスク・ブランクは、そこに意図的に設けたのかまたは無意識に設けたのかにかかわらず、ディスク・ブランクをさらに処理する前に、表面欠陥と特徴の存在について完全に分析することができる。したがって、欠陥のあるディスク・ブランクの高価な処理を回避することができる。
【００８３】
典型的な磁気データ記憶ディスクの熱背景信号は、５つの基本グループの周波数から構成されると見なすことができる。第１のグループは、最も主要なものであって、通常は２．５ＭＨｚから１０ＭＨｚの範囲になる、サーボ・パターン周波数を含む。第２のグループは、通常は６０ＫＨｚから７０ＫＨｚの範囲になる各サーボ・バーストの先頭から末尾までと定義されたサーボ長周波数を含む。第３のグループは、約１０ＫＨｚのサーボ間周波数またはサーボ・バースト間の時間の逆数を含む。第４のグループの周波数は、１０ＭＨｚを超えるデータ・パターン周波数である。トラックを消去する場合は、磁気データ・パターンを除去することができる。
【００８４】
第５のグループの周波数は、広帯域であり、ディスクの表面凹凸に関連し、ヘッド／ディスク間隔がディスクの表面変動の結果として変化する。第５のグループの上限は、ＭＲ応答の熱時間定数によって制限され、通常、それは約１マイクロ秒になる。適切なフィルタリングにより、この５つの「ノイズ」発生源の読戻し信号振幅変調の影響を選択的に抑制することができる。ただし、ヘッド／ディスク接触事象の信号対雑音比は、通常、１０：１（２０ｄｂ）を上回り、容易に検出できることに留意されたい。
【００８５】
上記の５つのノイズ発生源をフィルタリングするために、いくつかのフィルタリング方式を使用することができる。このようなフィルタリング方式としては、読戻し信号を除去するための楕円フィルタの使用を含む。帯域消去楕円フィルタにより、高い減衰が得られ、バッタワース・フィルタまたはチェビシェフ・フィルタより位相ひずみが小さくなる。有用なフィルタリング方式の１つでは、２つの楕円帯域消去フィルタを使用している。それぞれの４次ディジタル楕円帯域消去フィルタには、ｚ平面の単位円上に位置する伝達関数の２対の複素ゼロによる２つのノッチが付いている。一方の「低ノッチ」４次楕円フィルタにより、約１５ＫＨｚ以下の周波数が除去されるはずである。実際にサーボ間パターン周波数は、ディスク表面欠陥の所望の検出帯域幅に近いので、最も問題なものである。この低ノッチ楕円フィルタは、この周波数範囲で非常に高い減衰（たとえば、２０〜６０ｄｂ）を行う１２０Ｈｚと１０．８ＫＨｚのノッチをもたらすように設計することができる。高ノッチ・フィルタとして構成された第２の４次楕円ノッチ・フィルタは、約５ＭＨｚのサーボ・パターン周波数と約１５ＭＨｚのその第３調波を減衰するはずである。この２つの周波数は主要なものである。４次楕円ノッチ・フィルタは、この周波数範囲で非常に高い減衰を行うことができる。これらの周波数は、典型的なデータ記憶システムのスピンドルの速度が正確であるために不変である。
【００８６】
前述のように、読戻し信号の磁気成分と熱成分の両方は、そこから読戻し信号が読み取られたディスク表面の表面特性に関する情報を含む。様々なタイプのディスク表面欠陥と永続的に書き込まれたサーボ・セクタに対する磁気および熱間隔信号応答の比較については、以下の表１に示す。複合読戻し信号を独立した磁気信号成分と熱信号成分に分離することにより、ＭＲヘッドの２つの独立した同時応答を使用して同じ「未知の現象」または表面欠陥を検出する機会が得られる。ディスク表面分析中に２つの独立した熱信号と磁気信号を使用すると、欠陥検出解像度と信頼性の大幅な上昇が可能である。欠陥検出の機能強化は、１次元（１Ｄ）手法ではなく２次元（２Ｄ）検出手法を使用することによって実現される。１次元検出手法は、磁気信号と熱信号の両方ではなく、いずれか一方を使用する手法であると言われている。同じ瞬間に同じＭＲエレメントから得られる２つの独立した磁気信号と熱信号を使用することは、未知のディスク表面欠陥を検出し分類するための強力なツールになる。
【００８７】
図８に示す単純なディスク表面欠陥と永続記録に対する磁気および熱ＭＲヘッド／ディスク間隔応答間の差について、その概要を以下の表１に示す。掻き傷や打こんなど、多くのディスク表面欠陥は、通常、いくつかの単純な欠陥を複雑に組み合わせたものであり、そのため、より複雑なＭＲヘッド応答が得られることに留意されたい。
【００８８】
【表１】

【００８９】
上記の表１を参照すると、ディスク表面欠陥分析を実行するための方法は次のように行うことができる。第１に、データ記憶システム内に設けた各ディスクの各表面について熱走査を実行する。その結果得られる熱応答について、所定の正および負のしきい値を上回る熱電圧がないかどうか監視する。熱応答は磁気空隙と事前書込みしたサーボ・セクタに対して鈍感なので、このプロセスは、サーボ・セクタの磁気パターンと磁気空隙を有効な表面欠陥として除去することになる。ただし、磁気信号を排他的に使用する欠陥分析プロセスでは、磁気空隙や間違った位置の永続記録を検出したときに表面欠陥の存在を間違って示す恐れがあることに留意されたい。熱操作中の熱しきい値検出器の起動は、３つの基本タイプの表面欠陥、すなわち、ピット、こぶ、熱によるでこぼこ（ＴＡ）、あるいはこの３通りの欠陥タイプの組合せに帰因する可能性がある。
【００９０】
次に、上記の表１に示す磁気応答特性を使用して、ＨＲＦまたはＲＳＭ方法の使用などにより、磁気欠陥検証手順を実行するために熱しきい値が起動されたディスク表面位置に磁気情報を書き込むことができる。こぶと熱によるでこぼこだけが有効な障害条件を示すはずなので、その内部に欠陥ディスクが収容されているデータ記憶システムまたはディスクを排除する前に、熱検出器とＨＲＦ／ＲＳＭ検出器の同時起動を行わなければならない。
【００９１】
したがって、ヘッド／ディスク間隔の変化に応答するＭＲエレメントの加熱および冷却を使用すると、ディスクを排除する十分な原因になるはずのディスク２４に対する機械的損傷と、ディスクを排除する十分な原因にはならないはずのピットや磁気空隙などの他の致命的ではないディスク欠陥とを検出し、それらを区別することができる。このようなディスク表面欠陥分析は、現場でまたは出荷前に完全に動作可能なデータ記憶システム内で実行することができる。さらに、この分析は、データ記憶システムについて予測障害分析を実行するために、データ記憶システムの現場での耐用年数中に、所定の回数、現場で実行することができる。あるいは、欠陥のあるディスク・ブランクをさらに処理することを回避するために、磁気コーティングのないディスク・ブランクについてディスク表面欠陥分析を実行することもできる。
【００９２】
ディスク欠陥を分類する方法は、ＭＲヘッドから抽出した熱信号の負のピークと正のピークを使用して式で表すことができる。欠陥分類方法の一実施例は、ディスクのくぼみ（たとえば、ピット）の場合、ヘッド／ディスク間隔の増加のためにＭＲエレメントが加熱すると熱信号の振幅が増加するという判断から得られる。このクラスのディスク欠陥では、めったに冷却が発生しないので、ＭＲヘッドによって負の極性の熱信号が発生することになるはずである。
【００９３】
接近ヘッド／ディスク接触の条件の場合、ＭＲエレメントは冷える。ＭＲエレメントの冷却の結果、負の熱電圧信号Ｖ_THが発生する。機械的損傷の有無についてディスクをテストするための基準は次式で示される。
（Ｖ＋）＋｜Ｖ−｜＞Ｔ ［８］
式中、Ｖ＋はＶ_THの正のピークであり、Ｖ−はＶ_THの負のピークであり、Ｔはこれを超えた場合にディスクの機械的損傷の存在を示す熱電圧しきい値である。
【００９４】
式［８］のテスト基準を使用すると、既存のまたは切迫した機械的ディスク損傷の存在を正確に特定することができると発明者らは判断している。ディスクの機械的損傷の存在を判定する場合、式［８］の応用は妥当である。というのは、このような損傷は、ヘッドを変位させるようなディスク欠陥によって発生する加熱と、ディスク表面に近接しているためにＭＲエレメントによって発生する冷却の両方に関連するからである。めっきピットなど、上方へのディスク表面の突出に関連しないその他のディスク欠陥の場合、熱電圧信号は、懸念を保証するほど十分な大きさの負のピークを示さない。というのは、十分な量のＭＲエレメントの冷却が一切発生していないからである。さらに、磁気空隙など、磁気信号に大幅に影響するディスク欠陥は、適当な熱応答の発生に至らない。破滅的ではない磁気空隙を示す疑わしいディスクを排除し廃棄することは、これまでは磁気データ記憶ディスクのメーカによる一般的なやり方であった。というのは、従来のスクリーニング手順では、磁気空隙またはその他に関連するディスクの機械的損傷の有無を確実に検証することができないからである。
【００９５】
一般に、スクリーニングの合格基準と不合格基準とを区別するために曲線を使用することができる。このようなスクリーニング曲線の一例は、次式で示される。

_- ⁿ＋Ｃ₁Ｖ₊ ^m＝Ｃ₂ ［９］
式中、Ｖ_-は最小熱電圧であり、Ｖ₊は最大熱電圧であり、ｎ、ｍ、Ｃ₁、Ｃ₂は定数である。たとえば、ｎ＝ｍ＝２であり、Ｃ₁＝１である場合、合格不合格曲線は、半径が
【数７】
に相当する円の一部分になる。
【００９６】
特定のタイプの表面欠陥の識別は、ディスク表面欠陥検出回路を使用することによって確立することができる。図３１に示す実施例では、アナログ回路を使用して欠陥識別回路９１が実現されているが、欠陥識別回路９１はディジタル回路として実現するかまたはディジタル信号処理によって実現できることに留意されたい。欠陥検出識別回路９１は、たとえば、次式で示すように、負の熱冷却ピークＶ_TH（ｃｏｏｌ）と熱によるでこぼこの「加熱」スパイクの正のピークＶ_TH（ｗａｒｍ）との全熱電圧信号差ΔＶ_TH（ＴＡ）を測定することによって熱によるでこぼこ（ＴＡ）を検出する。
ΔＶ_TH（ＴＡ）＝Ｖ_TH（ｗａｒｍ）−Ｖ_TH（ｃｏｏｌ） ［１０］
【００９７】
熱電圧信号Ｖ_THは、信号分離／復元モジュール７６によって読戻し信号から抽出され、正のピーク保持回路７１と負のピーク保持回路７３に通信される。正のピーク保持回路７１は熱信号Ｖ_TH（ｗａｒｍ）の正のピーク電圧を緩衝し、負のピーク保持回路７３は熱信号Ｖ_TH（ｃｏｏｌ）の負のピーク電圧を緩衝する。比較器構成に演算増幅器を使用して実現したピット検出器７５は、適切な入力しきい電圧Ｔ_(PIT)と正のピーク電圧Ｖ_TH（ｗａｒｍ）とを比較することによって表面ピットを検出するように較正されている。
【００９８】
こぶ検出器７９は、同様に実現され、適切な入力しきい電圧Ｔ_(BUMP)と負のピーク電圧Ｖ_TH（ｃｏｏｌ）とを比較することによって表面のこぶを検出するように較正されている。また、熱によるでこぼこ検出器７７も同様に実現され、適切な入力しきい電圧Ｔ_(TA)と加算回路８１によって生成された熱差信号（Ｖ_TH（ｗａｒｍ）−Ｖ_TH（ｃｏｏｌ））とを比較することによって熱によるでこぼこを検出するように較正されている。正のピーク保持回路７１と負のピーク保持回路７３を使用しない実施例では、欠陥識別のために事前設定のしきい値に対して熱応答電圧Ｖ_THの最大正ピーク値と最小負ピーク値を連続監視してもよい。
【００９９】
３つの比較器７５、７７、７９の論理レベルは、以下の表２に示すように合格／不合格判断表と関連して使用する３ビット・ワード｛ＴＡ、ＢＵＭＰ、ＰＩＴ｝を形成するように構成することができる。前述のように、ディスク障害がある場合には、ディスクに磁気トラックを書き込み、ＨＲＦまたはＲＳＭあるいはその両方の検証テストを実行することによって、後で検証手順を実行することができる。
【０１００】
【表２】

【０１０１】
図３２を参照すると、同図には、読戻し信号の熱信号成分を使用するエラー回復ルーチンを実行するための様々なステップが示されている。図３２に流れ図形式で示したエラー回復ルーチンは、一般に、重大なエラー条件によるデータ回復のためには適当なものであり、通常は標準のエラー回復ルーチンをいくつか実行した後に行われる。ステップ５３０では、情報の紛失または読取り不能情報を含む欠陥セクタが識別される。ステップ５３２では、欠陥セクタまたはディスク域について読戻し信号がサンプリングされる。サンプリングした読戻し信号がステップ５３４で格納される。欠陥セクタに対応する格納読戻し信号は、ＲＳ_(DEFECT)＝Ｍ_(DEFECT)＋Ｔ_(DEFECT)として示される。ただし、ＲＳ_(DEFECT)は欠陥セクタから得られた全読戻し信号を表し、Ｍ_(DEFECT)は欠陥セクタから得られた読戻し信号の磁気信号成分を表し、Ｔ_(DEFECT)は欠陥セクタから得られた読戻し信号ＲＳ_(DEFECT)の熱信号成分を表す。
【０１０２】
ステップ５３６では、ディスク表面域の欠陥セクタが消去される。欠陥ディスク表面位置の消去が完了後、ステップ５３８に示すように消去した欠陥セクタについて読戻し信号がサンプリングされる。消去した欠陥セクタからの読戻し信号サンプルは、ＲＳ_(ERASE)＝Ｔ_(ERASE)として表される。ただし、ＲＳ_(ERASE)は消去した欠陥セクタからサンプリングした全読戻し信号を表し、Ｔ_(ERASE)は消去した欠陥セクタから得られた読戻し信号ＲＳ_(ERASE)の熱信号成分を表す。ただし、消去プロセスでは欠陥セクタから磁気信号成分のほぼすべてを除去してしまうので、読戻し信号ＲＳ_(ERASE)の磁気信号成分が含まれないことに留意されたい。
【０１０３】
しかし、ディスク表面上の微細亀裂は、消去手順にもかかわらず、少量の磁束を維持できることに留意されたい。このため、ステップ５４０に示すように、消去した欠陥セクタから得た読戻し信号から熱信号Ｔ_(ERASE)を抽出し格納することが望ましい場合がある。ただし、Ｔ_(DEFECT)は、図６および図７に関連して前述した熱信号波形によって検証されるようにＴ_(ERASE)とほぼ同等であることに留意されたい。ステップ５４２では、ステップ５３８で消去した欠陥セクタについて得られた読戻し信号がステップ５３２で欠陥セクタから得た読戻し信号から減算される。ステップ５４２に示すように、この減算により、欠陥セクタの回復磁気信号成分Ｍ_(RECOVERED)が得られる。次に回復磁気信号Ｍ_(RECOVERED)はメモリまたはデータ記憶ディスク上の他の場所に格納される。最後のステップ５４６では、回復磁気信号Ｍ_(RECOVERED)がメモリから通常のデータ記録チャネルを通過し、そこで情報が２進ワードに復号される。
【０１０４】
当然のことながら、本発明の範囲または精神を逸脱せずに、上記の実施例について様々な変更および追加を行うことができることに留意されたい。したがって、本発明の範囲は、上記の特定の実施例に限定すべきではなく、冒頭に記載した特許請求の範囲の完全かつ公正な範囲によってのみ規定すべきである。
【０１０５】
まとめとして、本発明の構成に関して以下の事項を開示する。
【０１０６】
（１）記憶媒体に近接している磁気抵抗（ＭＲ）エレメントを使用して記憶媒体から得た信号を処理するための方法において、前記方法が、
ＭＲエレメントを使用して記憶媒体から信号を読み取るステップと、
ＭＲエレメントの熱応答を表す熱信号成分をその信号から抽出するステップとを含むことを特徴とする方法。
（２）前記熱信号成分が、ＭＲエレメントと記憶媒体との距離を表すことを特徴とする、上記（１）に記載の方法。
（３）前記熱信号成分が、ＭＲエレメントと記憶媒体との距離の変動に応答して直線的に変動することを特徴とする、上記（２）に記載の方法。
（４）前記信号が、磁気信号成分を含み、
前記熱信号成分を使用してＭＲエレメントと記憶媒体との間隔の変動を推定するように、前記熱信号成分が磁気信号成分を使用して較正されることを特徴とする、上記（１）に記載の方法。
（５）前記信号が磁気信号成分を含み、前記方法が、
前記熱信号成分と磁気信号成分とを使用して熱間隔信号を発生し、熱間隔信号がＭＲエレメントと記憶媒体との間隔の変動に比例して変動するステップを含むことを特徴とする、上記（１）に記載の方法。
（６）前記熱間隔信号が、ＭＲエレメントと記憶媒体との間隔の変動に応じて直線的に変動することを特徴とする、上記（５）に記載の方法。
（７）前記熱信号成分が、記憶媒体の特性を表すことを特徴とする、上記（１）に記載の方法。
（８）前記記憶媒体の特性が、記憶媒体の表面輪郭であることを特徴とする、上記（７）に記載の方法。
（９）前記記憶媒体の特性が、記憶媒体の放射率であることを特徴とする、上記（７）に記載の方法。
（１０）前記抽出ステップが、前記信号から熱成分を除去するために信号をフィルタリングするステップを含むことを特徴とする、上記（１）に記載の方法。
（１１）前記抽出ステップが、有限インパルス応答（ＦＩＲ）フィルタを使用して前記信号をフィルタリングするステップを含むことを特徴とする、上記（１０）に記載の方法。
（１２）前記フィルタリング・ステップが、ＦＩＲフィルタに結合されたメモリに格納された１組のタップ重みを使用してＦＩＲフィルタをプログラミングするステップを含むことを特徴とする、上記（１１）に記載の方法。
（１３）前記フィルタリング・ステップが、ＦＩＲフィルタをプログラミングするときに前記１組のタップ重みにウィンドウ機能を適用するステップを含むことを特徴とする、上記（１２）に記載の方法。
（１４）前記フィルタリング・ステップが、
前記記憶媒体上に格納されたデータを読み取るために第１の組のタップ重みを使用してＦＩＲフィルタをプログラミングするステップと、
前記記憶媒体上に格納されたサーボ情報を読み取るために第２の組のタップ重みを使用してＦＩＲフィルタをプログラミングするステップとを含むことを特徴とする、上記（１２）に記載の方法。
（１５）前記信号がサーボ情報を含み、抽出した熱信号成分がサーボ情報を表すことを特徴とする、上記（１）に記載の方法。
（１６）前記信号が、サーボ情報を表す磁気信号成分を含むことを特徴とする、上記（１）に記載の方法。
（１７）前記信号から磁気成分を抽出するステップをさらに含むことを特徴とする、上記（１）に記載の方法。
（１８）前記信号が熱信号成分と磁気信号成分とを含み、前記方法が、
前記信号から抽出した熱信号成分を減算することにより、前記信号から磁気信号成分を抽出するステップをさらに含むことを特徴とする、上記（１）に記載の方法。
（１９）前記記憶媒体の欠陥部分から信号を読み取り、前記信号が磁気信号成分と熱信号成分とを含む複合信号であるステップと、
前記記憶媒体の欠陥部分から磁気信号成分を消去するステップと、
前記記憶媒体の消去済み欠陥部分から熱信号成分を抽出するステップと、
複合信号の磁気信号成分をほぼ表す復元磁気信号を生成するために、複合信号から抽出した熱信号成分を減算するステップとをさらに含むことを特徴とする、上記（１）に記載の方法。
（２０）情報記憶媒体を含む情報記憶装置用の信号分離装置において、
磁気抵抗（ＭＲ）エレメントを含む変換器と、
記憶媒体に近接したＭＲエレメントによって記憶媒体から信号を読み取るために変換器に接続された読取りチャネルと、
読取りチャネルに結合され、信号の熱信号成分を抽出するためのフィルタであって、熱信号成分がＭＲエレメントの熱応答を表すフィルタとを含むことを特徴とする装置。
（２１）前記フィルタが、有限インパルス応答（ＦＩＲ）フィルタであることを特徴とする、上記（２０）に記載の装置。
（２２）前記記憶媒体から読み取られた信号が、磁気信号成分を含むことを特徴とする、上記（２０）に記載の装置。
（２３）前記記憶媒体から読み取られた信号が、熱信号成分を含む磁気信号であることを特徴とする、上記（２０）に記載の装置。
（２４）前記記憶媒体から読み取られた信号が、サーボ情報信号であることを特徴とする、上記（２０）に記載の装置。
（２５）抽出した熱信号成分がサーボ情報信号であることを特徴とする、上記（２０）に記載の装置。
（２６）前記読取りチャネルに結合され、前記記憶媒体から読み取られた信号の磁気信号成分を抽出するための磁気信号フィルタを含むことを特徴とする、上記（２０）に記載の装置。
（２７）前記読取りチャネルとフィルタに結合された信号加算装置を含み、
信号加算装置が、抽出した前記熱信号成分と、前記熱信号成分と磁気信号成分とを含む前記記憶媒体から読み取られた複合信号とを受け取り、抽出した前記信号成分を使用して複合信号から熱信号成分を減算し、複合信号の磁気信号成分をほぼ表す復元磁気信号を生成することを特徴とする、上記（２０）に記載の装置。
（２８）情報記憶装置において、
磁気抵抗（ＭＲ）エレメントを含む変換器と、
記憶媒体と、
変換器と記憶媒体の少なくとも一方を移動させて、変換器と媒体との相対運動を行う手段であって、ギャップがＭＲエレメントを媒体から分離するように変換器が媒体に対して相対的に配置されている手段と、
ＭＲエレメントを使用して媒体から信号を読み取るために変換器に接続された読取りチャネルと、
読取りチャネルに結合され、信号の熱信号成分を抽出するためのフィルタであって、熱信号成分がＭＲエレメントの熱応答を表すフィルタとを含むことを特徴とする装置。
（２９）前記フィルタが、有限インパルス応答（ＦＩＲ）フィルタを含むことを特徴とする、上記（２８）に記載の装置。
（３０）ＦＩＲフィルタに結合されたメモリを含み、ＦＩＲフィルタがメモリに格納された１組のタップ重みを使用してプログラム可能であることを特徴とする、上記（２９）に記載の装置。
（３１）前記メモリが１組のウィンドウ・パラメータを格納し、
ＦＩＲフィルタが、前記メモリに格納された１組のタップ重みとウィンドウ・パラメータとを使用してプログラミングされることを特徴とする、上記（２９）に記載の装置。
（３２）前記ＭＲエレメントが前記媒体のデータ格納部分に近接して移動したときに第１の組のタップ重みがメモリからＦＩＲフィルタに転送され、
前記ＭＲエレメントが前記媒体のサーボ情報格納部分に近接して移動したときに第２の組のタップ重みがメモリからＦＩＲフィルタに転送されることを特徴とする、上記（２９）に記載の装置。
（３３）前記媒体から読み取られた信号が、磁気信号成分を含むことを特徴とする、上記（２８）に記載の装置。
（３４）前記媒体から読み取られた信号が、熱信号成分を含む磁気信号であることを特徴とする、上記（２８）に記載の装置。
（３５）前記媒体から読み取られた信号が、サーボ情報信号であることを特徴とする、上記（２８）に記載の装置。
（３６）抽出した熱信号成分がサーボ情報信号であることを特徴とする、上記（２８）に記載の装置。
（３７）前記読取りチャネルに結合され、前記媒体から読み取られた信号の磁気信号成分を抽出するための磁気信号フィルタを含むことを特徴とする、上記（２８）に記載の装置。
（３８）前記読取りチャネルとフィルタに結合された信号加算装置を含み、
信号加算装置が、抽出した前記熱信号成分と、前記熱信号成分と磁気信号成分とを含む前記媒体から読み取られた複合信号とを受け取り、抽出した前記信号成分を使用して複合信号から熱信号成分を減算し、複合信号の磁気信号成分をほぼ表す復元磁気信号を生成することを特徴とする、上記（２８）に記載の装置。
（３９）信号加算装置と書込みエレメントとを含み、
前記ＭＲエレメントが前記記憶媒体の欠陥部分から信号を読み取り、その信号が磁気信号成分と熱信号成分とを含む複合信号であり、
書込みエレメントが前記記憶媒体の欠陥部分から磁気信号成分を消去し、
前記フィルタが前記記憶媒体の消去済み欠陥部分から熱信号成分を抽出し、
信号加算装置が複合信号から抽出した熱信号成分を減算し、復元磁気信号を生成することを特徴とする、上記（２８）に記載の装置。
（４０）復元磁気信号が前記読取りチャネルで復号されることを特徴とする、上記（３９）に記載の装置。
（４１）媒体から間隔をおいた磁気抵抗（ＭＲ）エレメントで誘導された信号を使用して媒体を検査する方法において、
ＭＲエレメントが媒体に対して相対的に移動するときにＭＲエレメントを使用して信号を読み取るステップと、
ＭＲエレメントによって読み取られた信号の熱信号成分の変動を検出するステップと、
熱信号成分の変動から媒体の特性の変化を判定するステップとを含むことを特徴とする方法。
（４２）前記媒体の特性の変化が前記媒体の表面輪郭の変化であることを特徴とする、上記（４１）に記載の方法。
（４３）前記媒体の特性の変化が前記媒体の放射率の変化であることを特徴とする、上記（４１）に記載の方法。
（４４）前記検出ステップが、前記ＭＲエレメントと前記媒体との間隔の変化を検出するステップを含むことを特徴とする、上記（４１）に記載の方法。
（４５）前記媒体の一部分上の複数の位置で間隔を判定するステップと、
複数の位置で判定された間隔を使用して前記媒体の一部分の表面凹凸表現をマッピングするステップとを含むことを特徴とする、上記（４４）に記載の方法。（４６）前記媒体の一部分上の複数の位置で熱信号成分の変動を検出するステップと、
複数の位置での熱信号成分の変動を使用して前記媒体の一部分の表現をマッピングするステップとを含むことを特徴とする、上記（４４）に記載の方法。
（４７）前記媒体の特性が前記媒体の表面輪郭であり、前記方法が前記熱信号成分から熱間隔信号を生成するステップを含み、熱間隔信号が前記エレメントと前記媒体の表面との間隔の尺度を表すことを特徴とする、上記（４４）に記載の方法。
（４８）前記媒体の特性が前記媒体の表面輪郭であり、前記方法が、
前記ＭＲエレメントを使用して前記媒体から読み取られた信号の磁気信号成分の変動を検出するステップと、
磁気信号成分から磁気間隔信号を生成し、磁気間隔信号が前記ＭＲエレメントと前記媒体の表面との間隔の尺度を表すステップと、
熱信号成分から熱間隔信号を生成し、熱間隔信号が前記ＭＲエレメントと前記媒体の表面との間隔の尺度を表すように磁気間隔信号を使用して較正されるステップとを含むことを特徴とする、上記（４４）に記載の方法。
（４９）前記媒体の特性が前記媒体の表面輪郭であり、前記方法が、
熱信号成分の周波数を判定するステップと、
熱信号成分の周波数を前記媒体の表面輪郭特徴の１つに関連付けるステップとを含むことを特徴とする、上記（４４）に記載の方法。
（５０）前記方法が、前記ＭＲエレメントを使用して前記媒体から読み取られた信号の磁気信号成分の変動を検出するステップを含み、
前記媒体の特性が前記媒体の表面輪郭であり、
前記判定ステップが、熱信号成分の変動と磁気信号成分の変動とを使用して前記媒体の表面輪郭の変化を判定するステップを含むことを特徴とする、上記（４４）に記載の方法。
（５１）前記媒体の特性が前記媒体の表面輪郭であり、前記方法が、熱信号成分の変動を使用して表面輪郭の特徴を検出するステップを含むことを特徴とする、上記（４４）に記載の方法。
（５２）前記特徴の位置に対応する前記媒体上の位置に磁気信号を書き込むステップと、
前記ＭＲエレメントを使用して磁気信号を読み取るステップと、
前記ＭＲエレメントを使用して読み取られた磁気信号を使用して表面輪郭の特徴を検出するステップとを含むことを特徴とする、上記（５１）に記載の方法。（５３）前記媒体の特性が、前記媒体の表面上に設けられたくぼみを含む前記媒体の表面輪郭であり、前記方法が、
熱信号成分の変動を使用して前記媒体の表面上に設けられたくぼみを検出するステップを含むことを特徴とする、上記（４１）に記載の方法。
（５４）前記くぼみが前記媒体の表面上に設けられたグルーブを含むことを特徴とする、上記（５３）に記載の方法。
（５５）前記くぼみが前記媒体の表面上に設けられたピットを含むことを特徴とする、上記（５３）に記載の方法。
（５６）情報記憶装置において、
磁気抵抗（ＭＲ）エレメントを含む変換器と、
記憶媒体と、
変換器と記憶媒体の少なくとも一方を移動させて、変換器と媒体との相対運動を行う手段であって、間隔がＭＲエレメントを媒体から分離するように変換器が媒体に対して相対的に配置されている手段と、
ＭＲエレメントを使用して媒体から信号を読み取るために変換器に接続された読取りチャネルと、
読取りチャネルに結合され、信号の熱信号成分を通過させるためのフィルタであって、熱信号成分がＭＲエレメントの熱応答を表すフィルタと、
フィルタに結合され、記憶媒体の表面輪郭の変動に対応する熱信号成分の変動を検出するための検出器とを含むことを特徴とする装置。
（５７）前記表面輪郭の変動がサーボ情報を表すことを特徴とする、上記（５６）に記載の装置。
（５８）前記検出器によって検出された熱信号の変動が、前記ＭＲエレメントと前記記憶媒体の表面輪郭との間隔の寸法の変化に対応することを特徴とする、上記（５６）に記載の装置。
（５９）前記検出器が、前記記憶媒体の一部分上の複数の位置で間隔寸法の変動を検出し、前記記憶媒体の一部分の表面輪郭の変動を特徴づけることを特徴とする、上記（５８）に記載の装置。
（６０）前記検出器が、前記記憶媒体の一部分上の複数の位置で熱信号成分の変動を検出し、前記記憶媒体の一部分の表面輪郭の変動を特徴づけることを特徴とする、上記（５６）に記載の装置。
（６１）前記フィルタに結合され、熱信号成分を熱間隔信号に変換するための平均フィルタを含み、熱間隔信号が前記ＭＲエレメントと前記媒体の表面との間隔の寸法に対応することを特徴とする、上記（５６）に記載の装置。
（６２）前記フィルタとログ・フィルタとに結合され、前記ＭＲエレメントを使用して前記記憶媒体から読み取られた信号の磁気信号成分を磁気間隔信号に変換するための磁気信号フィルタを含み、磁気間隔信号が前記ＭＲエレメントと前記媒体の表面との間隔の寸法に対応することを特徴とする、上記（５６）に記載の装置。
（６３）前記フィルタとログ・フィルタとに結合され、前記ＭＲエレメントを使用して前記記憶媒体から読み取られた信号の磁気信号成分を磁気間隔信号に変換するための磁気信号フィルタを含み、磁気間隔信号と熱間隔信号が前記ＭＲエレメントと前記媒体の表面との間隔の寸法にほぼ対応することを特徴とする、上記（６２）に記載の装置。
（６４）熱信号成分の周波数を表面輪郭の特徴の１つに関連付けるための欠陥関連付け回路を含むことを特徴とする、上記（５６）に記載の装置。
（６５）前記検出器が、前記記憶媒体の表面輪郭の特徴に対応する、前記ＭＲエレメントを使用して前記記憶媒体から読み取られた信号の磁気信号成分の変動を検出し、
前記検出器が、表面輪郭の特徴に対応する、熱信号成分の変動を検出することを特徴とする、上記（５６）に記載の装置。
（６６）前記検出器が、熱信号成分の変動として前記記憶媒体の表面輪郭の特徴を検出することを特徴とする、上記（５６）に記載の装置。
（６７）変換器の単一磁気抵抗（ＭＲ）エレメントを使用して媒体から信号を読み取る方法において、ＭＲエレメントを使用して媒体から信号に含まれる第１のタイプの情報と第２のタイプの情報を同時に読み取るステップを含み、第１のタイプの情報が信号の磁気成分として表され、第２のタイプの情報が信号の熱成分として表されることを特徴とする方法。
（６８）前記第１のタイプの情報がデータを含み、前記第２のタイプの情報が変換器の位置情報を含むことを特徴とする、上記（６７）に記載の方法。
【図面の簡単な説明】
【図１】ＭＲヘッドで誘導された読戻し信号から熱信号および磁気信号を抽出するための装置のブロック図である。
【図２】その上部ハウジング・カバーが除去されたデータ記憶システムの平面図である。
【図３】読戻し信号の熱信号成分を獲得し使用するための方法を示す流れ図である。
【図４】ひずんだＤ．Ｃ．基線を示しているＭＲヘッドで誘導された読戻し信号を示す図である。
【図５】信号分離／変調モジュールによって処理した後の復元したＤ．Ｃ．基線を示している図４の読戻し信号を示す図である。
【図６】特定のトラック位置のＭＲエレメントで誘導された読戻し信号から抽出された熱信号を示す図である。
【図７】磁気情報のＡＣ消去後に同じトラック位置から得られる読戻し信号を示す図である。
【図８】様々な表面欠陥および特徴とこのような欠陥および特徴に対するＭＲエレメントの熱応答および磁気間隔応答とを示しているデータ記憶ディスクを示す拡大側面図である。
【図９】ヘッド／ディスク接触事象を示している読戻し信号を示す図である。
【図１０】ＭＲエレメントで誘導された読戻し信号から熱信号と磁気信号を抽出し、磁気信号のＤ．Ｃ．基線を復元するための信号分離／変調モジュールのブロック図である。
【図１１】磁気読戻し信号のＤ．Ｃ．基線を復元するための信号分離／変調モジュールのブロック図である。
【図１２】信号分離／変調モジュールに読戻し信号を選択的に通信するためのシステムのブロック図である。
【図１３】信号分離／復元モジュールで使用する有限インパルス応答（ＦＩＲ）フィルタの絶対値応答を示す図である。
【図１４】信号分離／復元モジュールで使用する有限インパルス応答（ＦＩＲ）フィルタの位相応答を示す図である。
【図１５】ＭＲヘッドで誘導された読戻し信号を示す図である。
【図１６】読戻し信号の復元磁気信号成分を示す図である。
【図１７】読戻し信号の未復元磁気信号成分を示す図である。
【図１８】信号分離／復元モジュールで使用するウィンドウ式ＦＩＲフィルタの絶対値応答を示す図である。
【図１９】信号分離／復元モジュールで使用するウィンドウ式ＦＩＲフィルタの位相応答を示す図である。
【図２０】典型的なＡＥモジュールの高域通過フィルタ挙動の絶対値応答を示す図である。
【図２１】典型的なＡＥモジュールの高域通過フィルタ挙動の位相応答を示す図である。
【図２２】典型的なＡＥモジュールとＡＥモジュールの有効高域通過フィルタのそれとは逆の伝達関数を有する逆フィルタとの高域通過フィルタ挙動の絶対値応答の比較を示す図である。
【図２３】典型的なＡＥモジュールとＡＥモジュールの有効高域通過フィルタのそれとは逆の伝達関数を有する逆フィルタとの高域通過フィルタ挙動の位相応答の比較を示す図である。
【図２４】図２３および図２４の逆フィルタを表す信号流れ図である。
【図２５】信号分離／復元モジュール内の処理点に発生したディスク表面ピットによる波形を示す図である。
【図２６】信号分離／復元モジュール内の処理点に発生したディスク表面ピットによる波形を示す図である。
【図２７】信号分離／復元モジュール内の処理点に発生したディスク表面ピットによる波形を示す図である。
【図２８】ディスク表面の欠陥に関連する磁気間隔信号と熱ヘッド／ディスク間隔信号との密接な対応を示す図である。
【図２９】無限インパルス応答（ＩＩＲ）フィルタを使用する信号分離／復元モジュールの他の実施例のブロック図である。
【図３０】ヘッド／ディスク接触事象に関連する磁気間隔信号と熱間隔信号を示す図である。
【図３１】欠陥分類回路のブロック図である。
【図３２】読戻し信号の熱信号を使用するエラー回復プロセスの流れ図である。
【符号の説明】
２４ データ記憶ディスク
６０ 読戻し信号
６１ａ 磁気信号成分
６３ 熱信号
７０ 装置
７２ 磁気抵抗（ＭＲ）エレメント
７４ ＡＥモジュール
７６ 信号分離／復元モジュール
７８ 磁気信号
８０ 熱信号[0001]
This application is a continuation-in-part of related US application Ser. No. 08/056164, filed Apr. 30, 1993.
[0002]
BACKGROUND OF THE INVENTION
The present invention relates generally to data storage systems, and more particularly to a method and apparatus for using the thermal component of a signal induced by a magnetoresistive (MR) head.
[0003]
[Prior art]
A typical data storage system includes a magnetic medium for storing data in magnetic form, and a transducer used to write magnetic data to the medium and read magnetic data from the medium. For example, a disk storage device includes one or more data storage disks that are coaxially attached to a spindle motor hub. The spindle motor usually rotates the disk at a speed of about several thousand revolutions per minute. Digital information is typically stored in the form of magnetic transitions on a series of concentric equally spaced tracks that form the surface of a magnetizable rigid data storage disk. Generally, a track is divided into a plurality of sectors, and each sector includes a plurality of information fields including fields for storing data, sector IDs, synchronization information, and the like.
[0004]
Actuator assemblies typically include a plurality of outwardly extending arms, with one or more transducers and a slider body attached to a flexible suspension. The slider body is typically an aerodynamic lift that lifts the transducer head off the surface of the disk as the spindle motor speed increases and causes the head to hover over the disk on the air bearing created by the high speed disk rotation. Designed as The spacing between the head and the disk surface is typically on the order of 50-100 nanometers (nm) and is commonly referred to as the head / disk spacing.
[0005]
In general, writing data to a data storage disk requires passing a current through the write element of the transducer assembly to generate magnetic flux lines that magnetize specific locations on the disk surface. Also, reading data from a specified disk location is typically accomplished by a transducer assembly read element that senses the magnetic field or flux lines emanating from the magnetized position of the disk. As the read element passes over the rotating disk surface, an electrical signal commonly referred to as a readback signal is generated in the read element as a result of the interaction between the read element and the magnetized position on the disk surface.
[0006]
Generally, conventional data storage systems use a closed loop servo control system to position the read / write transducer at a specified storage location on the data storage disk. During normal data storage system operation, a specified track is tracked (track tracking) using a servo transducer that is typically mounted near the read / write transducer or incorporated as a transducer read element. ) To read information for locating a specified track and data sector on the disk.
[0007]
According to one known servo technique, embedded servo pattern information is written to a disk along a segment extending in a direction generally outward from the center of the disk. Therefore, the embedded servo pattern is formed between the data storage sectors of the respective tracks. However, one servo sector is usually one piece of data used to maintain the optimum alignment of the read / write transducer on the track centerline when reading and writing data to a designated data sector on the track. Note that it includes a pattern (often called a servo burst pattern). The servo information can also include a sector and track identification code used to identify the position of the transducer.
[0008]
The data storage system manufacturing industry is currently interested in the use of MR elements as read transducers. MR heads typically incorporate MR read elements and thin film write elements and appear to have some advantages over conventional thin film heads, but the data storage system currently cannot accommodate some of the unwanted characteristics of MR heads. It is known to those skilled in the art that the advantages provided by MR heads are not completely realizable.
[0009]
In particular, MR element transducers typically introduce distortion into sensed magnetic signals representing data or servo information stored on a magnetic storage disk. The distortion of the magnetic signal is due to a number of factors, including some unwanted characteristics inherent in the MR element and the specific configuration and orientation of the MR element when incorporated into an MR transducer assembly. For example, it has been found that a typical MR element exhibits a variation in read sensitivity along the width of the MR element, which has been identified as one of the contributing factors of various severity servo control errors. Depending on the absolute value of the distortion of the magnetic signal caused by the MR element, the servo sector information is misunderstood or unreadable, resulting in interruption or loss of servo control, and in some cases the disk Unrecoverable loss of data stored above may occur.
[0010]
Considerable industry attention and resources have been and are still being spent to develop solutions aimed at reducing or eliminating the adverse effects associated with distorted magnetic readback signals. Such distortion of the readback signal obtained by the MR transducer has been treated as unwanted noise collectively, without fully evaluating the response of the MR element to various effects detected within its operating environment. . To date, no satisfactory solution has been found to eliminate or significantly reduce the distortion of the magnetic signal introduced by the MR element.
[0011]
[Problems to be solved by the invention]
In the data storage system manufacturing industry, there is an urgent need for an apparatus and method for eliminating unwanted distortion of magnetic readback signals induced by MR elements. There is also a need to provide such an apparatus and method suitable for incorporation into existing data storage systems as well as new system designs. The present invention is directed to such and other needs.
[0012]
[Means for Solving the Problems]
The present invention is an apparatus and method for reading an information signal from a magnetic storage medium using a magnetoresistive (MR) element and separating the thermal signal component and the magnetic signal component, if any, from the information signal. The magnetic signal is processed to remove the influence of the thermal signal component from the magnetic signal. The signal separation / restoration module eliminates the modulation of the magnetic signal component of the readback signal induced by the thermal signal component of the readback signal. A finite impulse response (FIR) filter can be used in the signal separation / restoration module to demodulate the magnetic signal. A thermal signal component can also be extracted from the readback signal using a signal separation / restoration module.
[0013]
According to a disk drive embodiment in which an MR element is coupled to an arm electronics (AE) module having high pass filter behavior, an inverse filter having a transfer function opposite to that of an effective high pass filter of the AE module is provided. The signal separation / restoration module can be configured to include. An infinite impulse response (IIR) filter can be programmed to reverse the amplitude and phase distortion of the thermal signal induced by the high-pass filter behavior of the AE module.
[0014]
According to another embodiment, the magnetic signal component and the thermal signal component of the readback signal are respectively extracted and processed so as to correspond linearly to the head / disk spacing. The head / disk spacing using the thermal signal can be used to detect disk surface defects and surface irregularities. The thermal signal can be calibrated using the magnetic spacing signal to directly measure changes in head / disk spacing. The thermal head / disk spacing signal can also be used for other systematic and diagnostic purposes, including defect characteristics, error correction, and predictive failure analysis.
[0015]
In other embodiments, servo control information can be provided on the disk surface in the form of surface irregularities variation. The thermal servo signal is induced at the MR element as the head passes over the surface irregularities variation and is extracted and communicated as a signal to the actuator and spindle servo controllers. Changes in the emissivity and / or absorption of the disk surface material can also be sensed by the MR element and converted into a corresponding thermal signal.
[0016]
DETAILED DESCRIPTION OF THE INVENTION
Referring now to the accompanying drawings, and more particularly to FIG. 1, the information signal having a magnetic signal component and a thermal signal component is read from the magnetic storage medium and the thermal signal component and the magnetic signal component are separated from the information signal. An apparatus 70 for the purpose is shown. The magnetic signal is processed to remove the influence of the thermal signal component from the magnetic signal. In this case, the use of two independent magnetic and thermal signals can increase the operation, performance and reliability of the data storage system.
[0017]
In FIG. 1, a magnetoresistive (MR) element 72 is shown proximate to the surface of the data storage disk 24. Generally, in this specification, information read from the disk 24 by the MR element 72 is called a read-back signal. The readback signal generated by the MR element 72 is typically amplified by an arm electronics (AE) module 74. Further, the read-back signal can be filtered by the AE module 74. As shown in the form of a graph on the output side of the AE module 74, the analog read-back signal 60 including the relatively high frequency magnetic signal component 61a has a D.D. C. The baseline distortion is shown. The modulated readback signal 60, or more specifically, the modulated magnetic signal component 61a of the readback signal 60, reduces the reliability of data storage and retrieval, such as servo control errors and inaccuracies, and in some cases cannot be recovered. It is understood by those skilled in the art that it has been identified as one source of the adverse effects of some data storage systems that can cause serious data loss.
[0018]
As previously noted in the prior art section, considerable interest and resources have been expended in the industry to fully understand the characteristics and origin of the baseline modulation of unwanted readback signals. As described in detail below, the inventors have discovered that the readback signal 60 is a composite signal that includes independent magnetic and thermal signal components, and the low frequency modulation of the readback signal is actually a readback signal. 60 independent thermal information signal components. As also detailed below, the inventors have further determined that unnecessary modulation of the readback signal 60 can be eliminated or the absolute value can be greatly reduced, resulting in data or servo It can correspond to a substantially pure magnetic signal representing information.
[0019]
The thermal signal component of the readback signal, which has previously been cumbersome, is generally referred to herein as the thermal signal and is extracted from the readback signal 60 and has various advantages that have not been recognized so far by those skilled in the art. Information content that can be used for various purposes. For example, the servo controller uses the thermal signal 63 to accommodate reliable track tracking and track seeking operations as opposed to using the magnetic signal 61a normally used according to conventional servo control techniques. be able to. In addition, the thermal signal 63 can be used to determine the flying height of the MR element 72 relative to the disk surface 24 to an accuracy of the order of 1 nanometer, including disk surface analysis and surface roughness mapping, disk defect detection and screening, error correction, It has been determined to include usable information for several other purposes, such as predictive failure analysis (PFA).
[0020]
The apparatus shown in FIG. 1 can be included as part of the design of a new generation data storage system that takes advantage of the availability of a thermal signal 63 that is independent of the restored unmodulated magnetic readback signal 61 and can be retrofitted to ) Can be incorporated into existing data storage systems that use standard MR heads. In general, as shown in FIG. 2, a data storage system 20 that uses an MR transducer typically includes one or more rigid data storage disks 24 that rotate about a spindle motor 26. The actuator assembly 10 typically includes a plurality of intervening actuator arms 11 and a suspension 12, each suspension supporting one or more MR head transducers 72 for reading and writing information to the data storage disk 24. is doing.
[0021]
Actuator assembly 10 includes a coil assembly 14 that cooperates with permanent magnet assembly 16 to operate as actuator voice coil motor 22 in response to control signals generated by controller 18. The controller 18 coordinates the transfer of data to and from the data storage disk 24 and moves the actuator arm 11 / suspension 12 and MR transducer 72 to the predetermined track 28 and sector 25 positions when reading and writing data on the disk 24. Cooperate with the actuator voice coil motor 22 to move.
[0022]
In order to understand the general process for separating the thermal and magnetic signal components from the readback signal as well as some useful applications that exploit the availability of the thermal information signal, reference is made to FIG. FIG. 4 illustrates in flowchart form a method for acquiring and using a thermal signal component of a readback signal according to one embodiment of the present invention. In step 30, information from the magnetic storage disk is received by MR element 72 so that MR element 72 includes a readback signal that may indicate distortion or modulation, but in fact includes a magnetic signal component and a thermal signal component. Read. However, it should be noted that the absolute value of the magnetic signal component is zero, and magnetic information may not be written to the disk at all. In general, however, there is a thermal signal component due to a previously unknown phenomenon now understood, characterized and used by the inventors. The thermal signal component is separated or extracted from the readback signal at step 32. The magnetic signal component of the read-back signal separated in step 32 is distorted by the amplitude of the magnetic signal component. C. Often includes baseline modulation.
[0023]
In step 34, the baseline modulation of the magnetic signal is removed, thereby restoring the baseline of the magnetic signal. If the extracted thermal signal is to be used in a servo controller, as tested in decision step 36, the thermal signal is transmitted to the servo controller in step 38 and processed accordingly. When the magnetic signal is used in the servo control device, the restored magnetic signal is transmitted to the servo control device in step 40. If the head float altitude routine needs to be executed as tested at decision step 42, a thermal signal is transmitted to the expected float altitude processor at step 44. If a disk surface analysis or screening needs to be performed, as tested at decision step 46, a thermal signal is transmitted to the disk surface analysis processor at step 48. Further, if the thermal signal is used to perform error recovery or predictive failure analysis (PFA) as tested in decision step 50, the thermal signal is transmitted to the error recovery processor at step 52.
[0024]
Turning now to FIGS. 4 and 5, a distorted readback signal and an undistorted readback signal restored by the signal separation / restoration module 76 as shown in FIG. 1 are shown. All low frequency distortion (modulation) can be corrected by the signal separation / recovery module 76, whether due to thermal head / disk spacing activity and / or unwanted filtering from the AE module 74. For purposes of illustration, assume that the readback signal 60 is a signal read from a servo sector on the data storage disk 24. The servo sector readback signal in this example includes a plurality of information fields: a write recovery field 62, a sync field 64, a gray code field 66, and a burst pattern field 68. Normally, the Gray code field 66 is understood to include a sector and cylinder identification field.
[0025]
In FIG. 4, it can be seen that the baseline of the servo-controlled readback signal is severely distorted, and particularly that distortion is severe within the gray code field 66. It should be noted, however, that the low frequency distortion associated with Gray code field 66 is not due to thermal head / disk spacing changes, but rather due to the high pass filter characteristics of AE module 74. Gray code field 66 is typically a detected amplitude that makes it difficult to reliably interpret sector and cylinder information in the presence of amplitude distortion. Such amplitude distortion often results in servo control errors of varying severity depending on the absolute value of the distortion. However, such amplitude distortion of the readback signal 60 obtained from the data sector, in contrast to the servo sector readback signal, results in detection and interpretation errors that result in soft or hard read errors. Note that there is a fear. It is not uncommon for up to 15-20% of the sectors formatted on the surface of disk 24 to exhibit a substantial level of readback signal distortion similar to that shown in FIG.
[0026]
As mentioned above, those skilled in the art have previously misunderstood that the source of unwanted modulation of the readback signal 60 is the noise of the MR element 72 or other sources of MR element 72 instability or malfunction. It was. However, the inventors have discovered that the readback signal distortion may be due to the presence of an independent thermal signal component of the readback signal that modulates the magnetic signal component 61a, thereby causing a readback signal having a time-varying baseline. 60 may occur. The signal separation / restoration module 76 processes the readback signal 60 to restore the readback signal baseline as shown in FIG. 5 and generates a substantially pure magnetic signal 61b that is not disturbed by the thermal signal component.
[0027]
The independence of the magnetic and thermal signals is demonstrated by the waveforms shown in FIGS. The waveform shown in FIG. 6 represents a thermal signal extracted from a magnetic readback signal using a digital filter configured as a low pass filter and an MR head. After acquiring the waveform shown in FIG. 6, magnetic AC erasure was performed on the track from which the waveform was generated. In order to obtain the waveform shown in FIG. 7, the same MR head was moved to the same position of the erased track. It can be seen that the extracted thermal signal shown in FIG. 6 and the read-back signal derived from the erased track shown in FIG. 7 are almost the same. The two waveforms shown in FIGS. 6 and 7 demonstrate that two simultaneously read thermal and magnetic signals are present in the readback signal and are independent and separable. .
[0028]
Looking at the readback signal with respect to its two independently separable components reveals information content that has not been previously recognized among the information content available in the readback signal obtained using the MR head. In particular, information about the surface of the disc can be derived from the thermal signal. An enlarged side view of the MR slider 67 and MR element 72 proximate to the surface 24a of the magnetic data storage disk 24 is shown in FIG. The disk surface 24a has surface irregularities that vary as a whole at the microscope level. As shown, the disk surface 24a may include various surface defects such as pits 122, humps 124, and surface portions 126 that are free of magnetic material. It should be noted that surface features such as grooves, pits, humps, etc. can be intentionally provided on the disk surface 24a to encode information on the disk surface 24a.
[0029]
As shown in FIG. 8, the thermal response voltage level 119 of the MR element 72 varies as a function of the spacing (indicated by parameter y) between the MR element 72 and the disk surface 24a. The change in the magnetic read back signal is due to the change in the resistance of the MR element 72. More specifically, a typical MR element is a resistor that is sensitive to the presence of a magnetic field and is electrically coupled to a current source between the positive and negative element leads. A bias current is applied to the MR element 72 through this lead. In normal operation, the magnetic transition on the disk surface 24 a affects the resistance of the MR element 72 and causes voltage fluctuations at the MR element 72. Such a voltage is generated at the frequency of the magnetic data transition recorded on the disk surface 24a and is the basis for the magnetic signal component of the readback signal.
[0030]
The resistance of the MR element 72 is also affected by the head / disk spacing. More specifically, as the head / disk spacing changes, a concomitant change in heat transfer from MR element 72 heated to a disk 24 by a constant bias current occurs. This heat transfer is an inverse function of the head / disk spacing. As heat transfer from the MR element 72 increases (small spacing), the temperature of the MR element 72 and its resistance decrease. As heat transfer is reduced, the temperature and resistance of the MR element 72 increase (increased spacing). Therefore, when the heat transfer between the MR element 72 and the disk 24 changes, the temperature of the MR element 72 changes. When the temperature of the MR element 72 changes, the electric resistance of the MR element 72 changes correspondingly, so that the voltage at the MR element 72 is supplied by a constant bias current. However, it should be noted that fluctuations in the flying height of the slider typically occur at frequencies significantly lower than that of the magnetic transition. Thus, such temperature changes in the MR element 72 occur at a much lower frequency than the magnetic data transition and are the basis for the thermal component of the readback signal.
[0031]
As shown in FIG. 8, there is an inverse relationship between the surface unevenness variation of the disk 24 and the change in the absolute value of the thermal signal 119. As the instantaneous head / disk spacing (y) increases, the air space insulation between MR element 72 and disk surface 24a increases correspondingly, thereby increasing the temperature of MR element 72. When the temperature of the MR element rises in this way, the resistance of the MR element 72 increases correspondingly because the temperature coefficient of the MR element material normally used for manufacturing the MR element 72 is positive. For example, permalloy is a preferred material used to fabricate the MR element 72, + 3 × 10^-3Shows a temperature coefficient of / ° C. As an example, as the MR element 72 passes over the hump 124 on the disk surface 24a, the heat transfer generated between the MR element 72 and the disk surface 24a increases, thereby cooling the MR element 72. When the MR element 72 is cooled in this manner, the resistance of the MR element decreases, and as a result, the voltage V at the MR element 72 with a corresponding constant bias current._THDecrease.
[0032]
As a result of the above interaction between MR element 72 and disk surface 24a, by referring to pits 122 on disk surface 24a, the thermal voltage at MR element 72 as a function of increasing head / disk spacing (y). Signal V_THIt can be seen that the amplitude of 119 increases. Further, by referring to the hump 124 on the disk surface 24a, the thermal voltage signal V as a function of the reduction in the head / di number spacing._THIt can be seen that the amplitude of 119 decreases. For convenience, the change in surface irregularities of the disk surface 24a is the thermal voltage signal V._THThe thermal voltage signal V so as to correspond directly to the change in 119, not the reverse._THIt may be desirable to invert 119. Therefore, the value of the MR head voltage is negative, that is, −V_THIn this case, the “cooling zone” is shown as a peak, and the “heating zone” is shown as a valley, whereby the surface irregularity of the disk surface 24a is qualitatively displayed.
[0033]
FIG. 8 also shows a magnetic interval signal 121 that has been conditioned to accommodate variations in the disk surface 24a. It can be seen that the magnetic spacing signal 121 falsely indicates the presence of some surface features, such as the magnetic air gap 126, as variations in surface irregularities on the disk surface 24a. Furthermore, it has been found that the magnetic spacing signal 121 indicates some other surface features, such as humps 124, when compared to the disk surface image information obtained through the use of the thermal signal 119.
[0034]
Note that in general the thermal signal includes a representation of the thermal response of the MR element as it interacts with the disk. If the emissivity or absorptance of the surface of the disk varies, the resulting thermal signal should vary accordingly. As will be better understood from the following description, the contents of both components of the readback signal are used to intentionally introduce disk surface variations such as surface contours or emissivity / absorbance. Can be used.
[0035]
Another characteristic of the MR element 72 that affects the characteristics of the readback signal obtained from the disk surface relates to situations where the MR element 72 is in physical contact with the disk surface or other obstacle. For example, when a temporary physical contact occurs between the disk surface and the MR element 72, a thermal asperity (TA) is generated due to heat. For example, the negative (cooling) peak of the thermal voltage response for the hump 124 is suddenly replaced by a large but narrow positive spike response that is followed quickly and continuously by a negative cooling response as shown in FIG. The positive spike response occurs due to mechanical frictional heat between the MR element 72 and local bumps on the disk surface 24a. The mechanical friction associated with heat bumps can scrape the magnetic coating in the area of physical contact. This causes the magnetic air gap 126, but is not the only source of such air gap.
[0036]
Referring to FIG. 10, there is shown an embodiment of the signal separation / restoration module 76 described above in connection with FIG. The signal separation / recovery module 76 performs a single task of separating the magnetic signal from the readback signal to remove the low frequency modulation component of the readback signal due to thermal signal effects or other causes. Note that you can. In another embodiment, the signal separation / restoration module 76 is used to separate the magnetic signal component from the readback signal 60 to remove the low frequency thermal signal component, and to extract the thermal signal from the readback signal, As a result, the dual task of making the information content of both magnetic and thermal signals available for subsequent processing in an almost independent manner can be performed.
[0037]
As shown in FIG. 10, the readback signal is induced by the MR element 72 located close to the magnetic data storage disk 24. As described in detail below, the frequency and amplitude of the readback signal modulation varies as a function of the behavior of the thermal signal component.
[0038]
In one embodiment, the readback signal received from the AE module 74 by the MR element 72 is converted from analog to digital form by the analog to digital converter 84. The digitized readback signal is then communicated to delay device 86 and linear phase programmable filter 88. Programmable filter 88 is a finite impulse response (FIR) filter having a length N, where N represents the number of impulse response coefficients or taps of programmable filter 88. The readback signal applied to the input of the programmable filter 88 is subjected to a total signal delay corresponding to the length N of the programmable filter 88 as it passes through the programmable filter 88.
[0039]
According to this embodiment, the programmable filter 88 is programmed with appropriate tap coefficients and weights to pass the relatively low frequency thermal signal component of the readback signal and remove the relatively high frequency magnetic signal component. The Thus, the programmable filter 88 is configured as a low-pass filter and is generally characterized as an intermediate frequency signal such that much of its energy falls in the frequency range from about 10 kilohertz (KHz) to about 100-200 KHz. Is programmed to pass. Note, however, that the magnetic signal component of the readback signal has a frequency in the range of about 20 megahertz (MHz) to 100 MHz. The thermal signal 80 on the output side of the programmable filter 88 is communicated to the signal adder 90. The thermal signal 80 can be transmitted from the output of the programmable filter 88 to other components in the data storage system, such as a servo controller, to control track tracking and track seek operations.
[0040]
Delay device 86 receives readback signal 60 from analog to digital converter 84 and provides signal adder 90 with a time interval equal to the delay time required for the readback signal to pass through programmable filter 88. Delay transmission of readback signal. Therefore, the readback signal including both the magnetic signal component and the thermal signal component and the thermal signal 80 extracted from the readback signal by the programmable filter 88 arrive at the signal adder 90 almost simultaneously. The signal adder 90 performs a demodulation operation of the readback signal and the thermal signal 80 and generates a restored readback signal 78. As a result, the signal separation / restoration module 76 illustrated in the embodiment shown in FIG. 10 can separate the magnetic signal component and the thermal signal component of the composite read-back signal, and further, restores the magnetic read-back without distortion. A signal 78 is generated.
[0041]
FIG. 11 illustrates another embodiment of the signal separation / restoration module 76 in which a restored magnetic readback signal 78 is generated after processing of the modulated readback signal by the signal separation / restoration module 76. According to this embodiment, an amplitude distortion readback signal including a magnetic signal component and a thermal signal component is sensed from the magnetic data storage disk 24 by the MR element 72 and communicated to the AE module 74. The modulated readback signal is then digitized by the sampler 84 and passed through a programmable filter 88 suitably configured to generate a distortion free restored magnetic readback signal 78. Programmable filter 88 is a finite impulse response (FIR) filter programmed to pass the relatively high frequency magnetic signal component of the composite readback signal and to eliminate the relatively low frequency thermal signal component of the composite readback signal. It is preferable that Although filters other than FIR filters can be used as the programmable filter 88, it is important that the filter 88 has a nearly perfect linear phase response to achieve optimal performance. This is easily accomplished using a digital FIR filter. However, depending on the application field, some degree of nonlinear phase behavior of the filter 88 may be allowed.
[0042]
In general, the signal is modulated in amplitude and / or phase as it passes through the filter. The characteristics and degree of modulation of the signal depend on the absolute value and phase characteristics of the filter. A filter's phase delay or group delay is a useful measure of how the filter modulates the phase characteristics of the signal. A filter with nonlinear phase characteristics introduces phase distortion in the signal passing through it. The reason for this phase distortion is that the frequency components of the signal are each delayed by an amount that is not proportional to the frequency, thereby changing the respective harmonic relationships. A given class of FIR filters removes almost all of the unwanted modulation of the readback signal resulting from the effects of the thermal signal component and provides the perfect linear phase response necessary to generate the restored magnetic readback signal 78. It is known that it can cope with.
[0043]
Referring to the embodiment shown in FIG. 12, the figure shows the function of selectively coupling and separating the signal separation / recovery module 76 from the recording channel through which the readback signal normally passes. Programmable filter 88 is shown coupled to a read only memory (ROM) 94 having a plurality of programmable filter parameter sets stored therein. In an embodiment using the FIR filter 88, the ROM 94 typically stores a plurality of tap weight sets 96, and may also store at least one restored tap weight set 98. As an example, for illustrative purposes, assume that a recording channel for a particular data storage system includes a single 10-tap FIR filter 88. The 10-tap FIR filter 88 is coupled to a ROM 94 configured to store 64 tap weight sets 96, any of which can be loaded into the FIR filter 88 to reprogram the response. As mentioned above, a modulated magnetic readback signal read from a servo sector can produce particularly detrimental results when the servo controller attempts to process a distorted readback signal. For example, sector and cylinder information contained in the servo sector gray code field 66 may be misunderstood or unreadable.
[0044]
The signal separation / restoration module 76 can be selectively used to process only readback signal information from servo sectors embedded between data sectors shown in FIG. According to this embodiment, the single programmable filter 88 used for the recording channel of the data storage system includes the signal separation / recovery module 76 and the read back signal processing corresponding to the servo sector by the servo channel, and the data It can be time shared with the processing of the readback signal corresponding to the data sector by the channel. When reading data sector information, the readback signal selectively passes through the data channel to bypass the signal separation / recovery module 76.
[0045]
As shown in FIG. 12, according to an embodiment of a data storage disk 24 using an embedded servo architecture, a series of alternating data sectors and servos are alternated when the data storage disk 24 rotates, typically at thousands of RPM. A sector passes under the MR element 72. When MR element 72 reads information from data sector 102, the readback signal generated at MR element 72 is routed to AE module 74, FIR filter 88, and data channel to bypass signal separation / restoration module 76. Is transmitted. Note, however, that FIR filter 88 is programmed with one of a plurality of tap weight sets 96 when processing signal information obtained from data sector 102.
[0046]
When the servo sector 104 is close to the MR element 72, the restored tap weight set 98 stored in the ROM 94 is loaded into the FIR filter 88, thereby replacing the previously loaded tap weight set resident in the FIR filter 88. . The restoration tap weight set 98 configures the FIR filter 88 to remove the thermal signal component of the read-back signal read from the servo sector 104, and restores the magnetic read-out corresponding to the pure magnetic signal stored in the servo sector 104. A return signal 78 is generated. The restored magnetic readback signal 78 is then communicated to the servo controller and processed accordingly. When the data sector 106 adjacent to the servo sector 104 is close to the MR element 72, the selected tap weight set 96 is loaded into the FIR filter 88 to replace the previously loaded restored tap weight set 98. Readback signals derived from the data sector 106 are processed by the FIR filter 88 and the data channel to bypass the signal separation / restoration module 76. The process of selectively processing the readback signal derived from the servo sector is repeated in a similar manner.
[0047]
It should be noted that the embodiment shown in FIG. 12 is particularly suitable for retrofitting a data storage system that includes a single programmable filter 88 in the read / write channel. It also incorporates an additional programmable filter so that the first programmable filter can be configured for operation on the servo channel and the second programmable filter can be configured for operation on the data channel. Note that it may be desirable. With such an arrangement using two independent programmable filters, restoring the readback signal derived from the data sector may improve the error rate performance of the data storage system.
[0048]
Note that the process of loading the restored tap weight set 98 stored in ROM 94 into the programmable filter 88 can be performed upon detecting a sync field 64 or other signal indicating the start of a servo sector. Similarly, sensing a sync field or other information signal indicating the start of a data sector determines when to load the programmable sector 88 with a tap weight set 96 for the data sector to read information from the data sector. can do. For details on designing, implementing, and programming FIR filters suitable for use in the signal separation / restoration module 76, see "Digital Signal Processing" by EC Ifeachor, BW Jervis (Addison-Wesley Publishing Company, Inc. 1993). I want.
[0049]
Returning to FIGS. 4 and 5, the modulated readback signal shown in FIG. 4 represents the appearance of the readback signal before it is processed by the signal separation / restoration module 76. The representation of the readback signal in FIG. 5 shows the readback signal in FIG. 4 after processing by the signal separation / restoration module 76. The unwanted effect of the thermal signal component on the readback signal shown in FIG. 4 is eliminated by using a 9-tap FIR filter in the signal separation / restoration module 76 to generate the restored magnetic readback signal 78 shown in FIG. ing. The absolute values and phase characteristics of the 9-tap FIR filter used to generate the restored magnetic readback signal 78 shown in FIG. 5 are shown in FIGS.
[0050]
In particular, it can be seen in FIG. 14 that the 9-tap filter exhibits a perfect linear phase response in the frequency range of interest. The effectiveness of the 9-tap FIR filter in removing baseline shift or modulation of the readback signal is demonstrated in FIGS. FIG. 15 shows a read-back signal that shows a baseline of unstable or amplitude variation. In FIG. 16, after passing the distorted readback signal through a properly programmed 9-tap FIR filter, the modulation baseline of the readback signal apparent in FIG. 15 is removed. The tap weights for the 9-tap filter used to restore the baseline of the readback signal are defined to include the following tap weights.
B (i) = (1/9) * (-1, -1, -1, -1, 8, -1, -1, -1, -1)
Or
B (i) = (-.111, -.111, -.111, -.111, .889, -.111, -.111, -.111, -.111)
[0051]
Note that the waveform shown in FIG. 17 was obtained by passing the modulated readback signal shown in FIG. 15 through a conventional high pass Butterworth filter. It can also be seen that there is still unnecessary modulation of the baseline of the readback signal even after passing the readback signal through a conventional high pass filter.
[0052]
As previously indicated, the absolute value and phase characteristics of the 9-tap FIR filter used to restore the baseline of the readback signal shown in FIG. 16 are shown in FIGS. 13 and 14, respectively. In FIG. 13, it can be seen that a certain amount of ripple that can be removed by applying the window function to the tap weight of the 9-tap FIR filter may occur in the passband of the filter. As an example, a Hamming window can be applied to the tap weights of a 9-tap FIR filter to create a windowed restoration filter with the following tap weights.

As a result of the output of the 9-tap window FIR filter having the above tap weights, ripples are removed as shown in FIG. As further shown in FIG. 19, the windowed 9-tap FIR filter retains its perfect linear phase response. Note, however, that applying a window function such as a Hamming window to the tap weights of the programmable FIR filter 88 allows for non-zero DC gain and some increase in low frequency response.
[0053]
Turning now to FIGS. 20-30, a signal separation / reconstruction module 76 that is particularly suitable for data storage systems that use high-pass filtering of readback signals acquired from the magnetic data storage disk 24 using the MR element 72. Another embodiment of is shown. This embodiment can be used in newly designed data storage systems as well as retrofit systems. In the design of an analog AE module 74 such as the AE module 74 shown in FIG. 1, it is desirable to include high-pass filtering with a preamplifier to eliminate the signal content of the readback signal below the frequency range of the magnetic signal component. There are many. Due to the high-pass filtering behavior of the AE module 74, the thermal signal component of the composite readback signal is distorted in both amplitude and phase. This thermal signal distortion varies in severity depending on the frequency and phase response of the particular AE module used.
[0054]
As an example, a high pass filter suitable for use with the AE module 74 can have a cutoff frequency of about 500 KHz and exhibit nonlinear phase behavior. However, the frequency associated with meaningful thermal signal information is typically 200 KHz or less, ranging from 10 KHz to about 100 KHz. Note that a high pass filter having a cutoff frequency of about 500 KHz significantly distorts the amplitude and phase of the thermal signal component of the readback signal. However, the magnetic signal component of the readback signal remains unaffected by the high pass filter. This is because the frequency range of the magnetic signal is generally about 20 to 40 times the high-pass filter cutoff frequency.
[0055]
FIGS. 20 and 21 show graphs showing the absolute value and phase response of a typical analog AE module 74 showing high-pass filtering behavior, respectively. This high-pass filter has a cutoff frequency of about 500 KHz. The digital version equivalent to the analog transfer function of the effective high pass filter of the AE module 74 having the absolute value and phase response shown in FIGS. 20 and 21 and a single pole of 500 KHz can be defined as follows.
[Expression 1]

■ F
b_h(1) =. 9876
b_h(2) =-. 9876
a_h(2) =-. 9752
[0056]
The absolute value and phase distortion of the thermal signal caused by the high pass filtering behavior of the AE module 74 is effectively eliminated by the use of an inverse filter having a transfer function opposite to that of the high pass filter. When the readback signal output from the AE module 74 is passed through an inverse filter, the thermal signal is restored to its original form in both amplitude and phase. For example, the transfer function of the inverse filter for adjusting the condition of the read-back signal that has passed through the high-pass filter having the above-described transfer function of Expression [1] is expressed by
[Expression 2]

O057]
The absolute value and the phase response of the effective high-pass filter of the AE module 74 and the inverse filter described in Equation [2] are shown in FIGS. 22 and 23, respectively. In particular, the absolute value responses of the inverse filter and the effective high pass filter of the AE module 74 are shown as curves 170 and 172, respectively, in FIG. Also, the phase responses of the inverse filter and the effective high-pass filter are shown as curves 176 and 174 in FIG. 23, respectively.
[0058]
In one embodiment, an infinite impulse response (IIR) filter is used as an inverse filter in the signal separation / restoration module 76 to recover the thermal signal content of the high pass filtered readback signal. The impulse response of an IIR filter has an infinite duration in contrast to an FIR filter where the impulse response has a finite duration. Also, unlike FIR filters, which may exhibit a perfect linear phase response, the phase response of IIR filters is non-linear, especially at the band edges. In an alternative embodiment, an analog filter can be used, but the IIR filter can be used as an inverse filter to restore the amplitude and phase of the thermal signal distorted by the high-pass filtering behavior of the analog AE module 74. Providing several suitable advantages.
[0059]
The signal flow diagram shown in FIG. 24 represents a first order IIR filter configured as an inverse filter. The coefficients associated with the signal flow diagram of FIG. 24 for a first-order IIR inverse filter having the transfer function shown in equation [2] above are as follows:
a₁=. 9876
a₂=-. 9876
b₁=. 1
b₂=-. 9752
For details on the design, implementation and programming of IIR filters suitable for use as inverse filters in the signal separation / restoration module 76, see “Digital Signal Processing” by EC Ifeachor, BW Jervis (Addison-Wesley Publishing Company, Inc. 1993). Refer to).
[0060]
FIGS. 25-27 show three waveforms used to demonstrate the effectiveness of the inverse filter to restore the original amplitude and phase of the thermal signal component of the readback signal that passed through the high pass filter. It is shown. FIG. 25 shows the read-back signal detected by scanning the pits on the data storage disk surface using the MR head. The magnetic read back signal shown in FIG. 25 is detected from a track written at a writing frequency of 20 MHz. This magnetic read-back signal is sampled at 100 MHz with 8-bit resolution. The signal shown in FIG. 26 represents the absolute value between the calculated peaks of the readback signal in FIG. In addition, the signal shown in FIG. 26 represents the magnetic signal component of the readback signal shown in FIG. 25 that clearly demonstrates a significant reduction in amplitude as the MR read element passes over the pits. FIG. 27 shows the thermal signal component of the readback signal of FIG. 25 after passing through the effective high-pass filter of the AE module 74. Comparing the waveforms of FIGS. 26 and 27, it can be seen that the magnetic signal component and the thermal signal component of the readback signal do not correspond closely to each other. The fact that the thermal signal and magnetic signal do not correspond so much is a result of distortion of the thermal signal due to the effective high-pass filtering characteristics of the analog AE module 74, thereby effectively distinguishing the thermal signal. .
[0061]
The inverse filter of the signal separation / restoration module 76 restores the amplitude and phase of the thermal signal 162 as shown in FIG. However, it should be noted that the thermal and magnetic signals shown in FIG. 28 are shown as head / disk spacing signals as will be described in detail below. It can be seen that the magnetic signal and the restored thermal signal show a close correspondence with each other after the high pass filtered thermal signal has passed through the inverse filter.
[0062]
As described above in connection with FIG. 8, the thermal signal induced in the MR head varies as a function of the head / disk spacing. Therefore, if the information contained in the heat signal is used, it is possible to detect fluctuations in the surface roughness of the disk. The thermal signal can be used to detect various surface features such as pits, bumps, bumps, heat bumps, and contaminating particles. Note that such surface features can be intentionally incorporated into the disk surface to derive various types of information using thermal signals.
[0063]
For example, concentric and radially indentations can be included in the disk surface to determine track and sector positions using thermal signals. A detailed surface roughness mapping of the disk surface can also be performed using the thermal signal. Note that the availability of the thermal signal extracted from the readback signal can be exploited for use in a wide range of applications. As another example, one or more indentations are formed in the disk surface to a known depth to calibrate the thermal response of the MR head in order to determine the head / disk spacing measurement using the thermal signal. You can also
[0064]
It is known to those skilled in the art to use the magnetic read signal generated by the read / write transducer to determine the change in spacing between the disk surface and the transducer. One such method for determining head / disk spacing using a magnetic readback signal is referred to as a harmonic ratio floating height (HRF) clearance test. The HRF test is a known method for measuring the flying height of a slider supporting a transducer, performed in the field or in a data storage system housing using a magnetic head / disk spacing signal. The HRF method is described in US Pat. No. 4,777,544, assigned to the assignee of the present invention and incorporated herein by reference. This HRF measurement method is a method in which the ratio of two spectral lines in the spectrum of the read-back signal is continuously and instantaneously measured. The amplitudes of both instantaneous spectral lines are related to the same volume element of the recording medium immediately below the MR transducer. This HRF measurement method can also determine the instantaneous head clearance relative to the disk surface using a magnetic readback signal.
[0065]
According to one embodiment, the thermal signal component of the readback signal induced by the MR head is used to qualitatively determine the change in head / disk spacing. In another embodiment, the thermal signal is calibrated using a magnetic signal to accommodate a quantitative determination of head / disk spacing. Referring to FIG. 29, there is shown in block diagram form a system for processing readback signals to obtain magnetic and thermal head / disk spacing information. The read back signal is detected from the disk surface 24 by the MR element 72. For purposes of illustration, assume that the readback signal is a composite signal that includes both a magnetic signal component and a thermal signal component, and a readback signal lacking the magnetic signal component is useful for determining head-to-disk clearance. It is understood to include a thermal signal component. The readback signal detected by the MR element 72 is communicated to the AE module 74 and then to the high pass filter 150. The high pass filter 150 is shown as a component external to the AE module 74, but is provided to generally represent the high pass filtering behavior of the AE module 74. The transfer function of the effective high-pass filter 150 is H₀Is shown as The output signal from the high pass filter 150 is sampled by an analog to digital converter 151 to generate digitized samples of the high pass filtered readback signal.
[0066]
As shown in FIG. 29, the thermal signal at point 159 on the output side of thermal signal extraction filter 157 can be generated using any of the methods described above. For example, the digitized readback signal can be communicated to an inverse filter 156 that corrects for distortion caused by the high pass filter 150 of the AE module 74. The transfer function of the inverse filter 156 is H₀ ^-1Is shown as The thermal signal is then extracted by thermal signal extraction filter 157, which can be a FIR filter. Note that the inverse filter 156 and the thermal signal extraction filter 157 can be implemented in a single IIR filter to recover the thermal signal distorted by the high pass filter 150. Alternatively, the readback signal can be tapped at some point before the high pass filter 150 and input to the thermal signal extraction filter 157, which can be an FIR filter as detailed above. . The thermal signal extracted by the thermal signal extraction filter 157 is communicated to an average filter 158 which in turn generates a thermal spacing signal 162 that is linearly related to the head / disk spacing. Average filter 158 is a digital motion smoothing averaging filter.
[0067]
The readback signal output to the output side of the analog / digital converter 151 is also detected by an amplitude detector 152 such as an FIR filter that detects the peak-to-peak amplitude of the readback signal and extracts a magnetic signal component from the readback signal. Can communicate. The logarithm of the magnetic signal is obtained by passing the magnetic signal through the log device 154, which generates a magnetic signal that is linearly related to the head / disk spacing. Once both the magnetic spacing signal 160 and the thermal spacing signal 162 are extracted, the thermal signal can be calibrated. This is because magnetic calibration is known and depends only on the recorded wavelength of the signal. It is important to note that the negative (or reciprocal) of the magnetic spacing signal 160 and the extracted thermal signal 162 are linearly proportional to the head / disk spacing (y).
[0068]
In FIG. 28, the thermal interval signal 162 processed by the thermal signal extraction filter 157 and the average filter 158 is shown along with the magnetic interval signal 160 processed by the amplitude detector 152 and the log device 154. It should be noted, however, that the linearized magnetic spacing signal 160 is usually calculated by taking the logarithm of the peak-to-peak signal and multiplied by the known sensitivity of the output voltage change to the magnetic spacing change by the well-known Wallace equation. Also, in FIG. 28, it can be seen that the magnetic spacing signal 160 and the thermal spacing signal 162 represent pits on the disk surface, except for the difference in signal height associated with the thermal spacing signal 162 and a slightly longer time constant. As detailed below, the linearized magnetic spacing signal 160 can be used to calibrate the thermal spacing signal 162 to accurately reflect the true head / disk spacing.
[0069]
An important advantage of the present invention relates to the ability to use the thermal response of the MR element 72 to detect changes in head / disk spacing in the field or in the housing of the data storage system. Field head / disk spacing measurement using the thermal response of MR element 72 is for disk manufacturing testing and screening, and to perform predictive failure analysis (PFA) during the life of the data storage system in the field. Useful. In addition, the thermal interval signal 162 can be used to detect head contact with the surface of the data storage disk.
[0070]
Referring now to FIG. 30, both the magnetic spacing signal 160 and the thermal spacing signal 162 are shown for a head / disk contact event, such as contact between an MR head and a local thermal bump (TA). The magnetic spacing signal 160 is linearized by taking the logarithm of the magnetic signal. Also, the thermal interval signal 162 is determined by using the inverse filtering technique described above. It can be seen that the MR element / disk spacing increases when the MR element 72 is displaced upward from the surface of the disk due to the unevenness of the disk. Both the magnetic spacing signal 160 and the thermal spacing signal 162 indicate that the head / disk spacing gradually increases in this manner from 0 microseconds to about 25 microseconds. After the MR element 72 passes over the bumps, a certain amount of air bearing (head / disk spacing) modulation occurs before the MR element 72 returns to its steady state flying height. It can be seen in FIG. 30 that this air bearing modulation starts at about 35 microseconds and continues to 70 microseconds.
[0071]
The similarities in the waveform characteristics of the magnetic spacing signal 160 and the thermal spacing signal 162 demonstrate that the thermal spacing signal 162 can be used to detect head / disk contact in the field without resorting to test bench equipment or an external tester. ing. Note, however, that the inverse filter characteristics required for a particular data storage system depend on the pole position of the high pass filter 150 of the AE module 74. For embodiments that use IIR or FIR filters, only the coefficients or tap weights need to be changed. In embodiments that use IIR filters, this change can be made adaptively or dynamically when variations occur in the pole frequency of the high pass filter 150. Usually, such fluctuations occur with changes in temperature or the like. However, it should be noted that the inverse filter 150 described herein is not limited to a first order IIR structure.
[0072]
In general, the absolute value of the thermal signal induced by the MR head is a function of the particular MR element used in the MR head. For example, variations in the manufacturing process and materials used can cause variations in MR element response. Therefore, it is desirable to calibrate the thermal response in situ using the magnetic response in order to accurately determine head / disk spacing changes using the thermal response of the MR head. For example, accurate magnetic spacing information can be determined using the well-known Wallace spacing loss equation. The use of calibration indentations such as radial trenches or pits in the landing zone can generate both thermal and magnetic signal modulations to perform on-site thermal spacing calibration. For that trench, the magnetic spacing can be accurately determined and can be used to calibrate the thermal voltage response of the MR element.
[0073]
Another method involves combining magnetic head / disk spacing measurements obtained using the HRF method or other similar methods with thermal clearance measurements. According to this coupling test, thermal and magnetic simultaneous “spin-down” is performed, whereby the thermal voltage change between two disk velocities is compared to the known (HRF) spacing change between the two disk velocities. . However, recovery of the thermal signal at disk rotational speed may be difficult to achieve in a system that uses an AE module 74 that has high pass filtering behavior. This is because the high-pass cutoff frequency is usually several orders of magnitude greater than the disk rotation frequency.
[0074]
The head / disk spacing calibration of MR head thermal response is made more complicated by the high pass frequency filtering characteristics of the AE module 74. Transfer function H of AE module 74 with high pass filtering behavior_AEIn general, (s) can be expressed up to a first order approximation as follows.
[Equation 3]

Where K_AEIs the gain of the AE module 74 at the recording frequency, and “a” is the cutoff frequency for the effective high-pass filter incorporated in the AE module 74. Typical values for gain are K_AE= 170, but the gain of the AE module 74 generally varies greatly. A typical cutoff frequency “a” is about 325 KHz, associated with a large tolerance of generally +/− 125 KHz.
[0075]
The frequency of interest for hump-like surface defects detected during surface analysis screening is typically in the range of 10 KHz to 100 KHz. The thermal response of the MR head directly converts such frequencies, while the magnetic response shifts these frequencies to the 20 MHz range due to the magnetic recording carrier frequency. The high pass characteristic of the AE module 74 attenuates all hump disturbance amplitudes in the thermal response by a variable amount for frequencies below 400 KHz, but the magnetic response is unaffected. Some form of integration needs to be applied to restore the decay of the thermal response. This restoration process is related to the transfer function H of the AE module 74._AETransfer function H opposite to (s)_INVCan be implemented by using an inverse filter with (s) (ie, H_INV(S) = 1 / H_AE(S)). For example, the lowest frequency from an MR head that reads data from a disk rotating at 7200 RPM is 120 Hz, and the lowest frequency to detect a hump on the disk surface is about 10 KHz. Therefore, pseudo-inverse filters such as advance / lag filters that are zero (“a”) at 400 KHz and pole (“b”) at 5 KHz may be more suitable for this application.
[0076]
The pseudo inverse filter appears to have a transfer function of the form
[Expression 4]

The overall transfer function of the pseudo inverse filter cascaded with the AE module 74 is as follows.

Thus, the corrected transfer function H (s) is a high pass filter with a cut-off frequency of 5 KHz, which is reasonable to pass the frequencies associated with the disk surface bumps in an undistorted form. It is.
[0077]
Since the fluctuation of the high-pass cutoff frequency “a” is large, a large fluctuation may occur in the restored thermal response. High pass cutoff frequency “a” and gain K of each MR head_AEAccurate estimation of sensitivity [nm / mv] is important for reliable calibration. Due to the lack of low frequency (ie, 120 Hz or less) response of the AE module 74, this thermal calibration process may be supported by another method as a reference standard. This support method, referred to herein as a magnetic readback signal modulation (RSM) method, is a known self-calibration method for determining head / disk magnetic spacing changes based on Wallace spacing loss techniques. An effective thermal calibration procedure is based in part on the RSM method, where the actuator is used to stop the crash in the landing zone to inspect both the thermal and magnetic components of the track readback signal, including the calibration indentation. It is executed first while being pressed.
[0078]
Polishing and sputtering processes are applied to obtain both gas and thermal data. However, it should be noted that the depression formed in the landing zone can also be formed as a manufactured bump. However, the hump is more likely to generate head / disk interference (HDI) and may cause permanent damage to the head or disk as a result of a head / disk crash. Also, the hump may be unsuitable for use as a permanent calibration location because it can cause head lift and air bearing modulation. In contrast, “pure” pits do not cause head lift or air bearing modulation. The calibration trench on the disk substrate surface is also inexpensive to manufacture.
[0079]
Before describing one embodiment of the configuration procedure, it may be useful to define some variables related to the calibration process. However, it should be noted that the term LF (low frequency) refers to a frequency on the order of the disk rotational frequency (RPM / 60), such as 120 Hz when the rotational speed is 7200 RPM. Referring to Equation 6 and Equation 7, V_THThe term (LF), excluding data obtained from calibration pits, represents the average per landing rotation of the AE module restoration thermal voltage (baseline) response in the landing zone and is usually expressed in millivolts (mv). V_THThe term (Pit) is the AE module restored average thermal voltage peak generated from the landing zone calibration pits taken for multiple revolutions and is usually expressed in millivolts (mv). δ_HRFThe term (LF) represents the average estimate per revolution of the RSM head / disk spacing in the landing zone, excluding data obtained from calibration pits, and is usually expressed in nanometers (nm). Finally, δ_HRFThe term (Pit) is the average peak HRF head / disk spacing generated from the landing zone calibration pits taken for multiple rotations, usually expressed in nanometers (nm).
[0080]
The proposed thermal calibration process uses an average low frequency (LF) RSM estimate δ per rotation of the head / disk spacing._HRF(LF) and the corresponding average thermal baseline voltage V_THBased on the use of (LF), these are jointly obtained with the actuator leaning against the landing zone crash stop, excluding data obtained from the calibration pits. By the average peak value of the excluded heat and magnetic pit data, V_TH(Pit) and δ_HRF(Pit) is obtained.
[0081]

The head “AC” calibration factor C (i) can be determined from:
[Formula 6]
For pits, δ_HRF(LF)> δ_HRF(Pit) and δ_TH(LF)> δ_THIt should be understood that the condition (Pit) is true. An approximate expression of the i-th thermal head / disk spacing is as follows.
δ_TH(I) = δ_HRF(LF) + C (i) · ΔV_TH                  [7]
Where ΔV_TH= V_TH(Defect) -V_TH(LF). In the case of a hump, it is assumed that there is no head / MR element contact and there is cooling, so ΔV_TH<0. In the case of pits, head / disk separation increases and MR element heating occurs, thus ΔV_TH> 0. The thermal head / disk spacing approximation is calibrated at the position of the calibration pit (ie, this is the inner diameter landing zone, but should be the outer diameter for load / unload disk drives) Therefore, the average RSM head / disk spacing δ at the track radius where the defect occurs_HRFBy updating (LF), an improvement in accuracy can be realized. In order to do this during manufacturing screening, a magnetic track should be written to the defect radius.
[0082]
Even if the thermal response of the MR head is not calibrated, the thermal signal component of the read-back signal can be used to perform a qualitative analysis rather than a quantitative analysis of the disk surface characteristics. Thus, disk surface analysis screening can be performed to detect disk defects using a unique normalization technique. One such approach is based on using the unique “background” thermal signal information on the disk track as a basis from which the clip level (ie failure threshold) is derived. . It is important to note that both quantitative and qualitative evaluation of the disk surface irregularities can be performed without applying a magnetic coating to the disk surface. Thus, a disk blank without a magnetic coating, whether intentionally or unintentionally provided there, is fully analyzed for the presence of surface defects and features before further processing of the disk blank. be able to. Thus, expensive processing of defective disk blanks can be avoided.
[0083]
The thermal background signal of a typical magnetic data storage disk can be considered to be composed of five basic groups of frequencies. The first group includes the servo pattern frequencies, which are the most dominant and usually range from 2.5 MHz to 10 MHz. The second group includes servo length frequencies defined from the beginning to the end of each servo burst, typically in the range of 60 KHz to 70 KHz. The third group includes the reciprocal of the inter-servo frequency or servo burst time of about 10 KHz. The fourth group of frequencies is a data pattern frequency exceeding 10 MHz. When erasing a track, the magnetic data pattern can be removed.
[0084]
The fifth group of frequencies is broadband and is related to disk surface irregularities, with the head / disk spacing changing as a result of disk surface variations. The upper limit of the fifth group is limited by the thermal time constant of the MR response, which is usually about 1 microsecond. By appropriate filtering, the influence of the read-back signal amplitude modulation of these five “noise” sources can be selectively suppressed. However, it should be noted that the signal to noise ratio of head / disk contact events is typically greater than 10: 1 (20 db) and can be easily detected.
[0085]
Several filtering schemes can be used to filter the above five noise sources. Such a filtering scheme includes the use of an elliptic filter to remove the readback signal. Band elimination elliptic filters provide high attenuation and less phase distortion than Butterworth or Chebyshev filters. One useful filtering scheme uses two elliptical band elimination filters. Each fourth-order digital elliptical band elimination filter has two notches with two pairs of complex zeros of the transfer function located on the z-plane unit circle. One “low notch” fourth order elliptic filter should remove frequencies below about 15 KHz. Actually, the pattern frequency between servos is the most problematic because it is close to the desired detection bandwidth of the disk surface defect. This low notch elliptic filter can be designed to provide 120 Hz and 10.8 KHz notches with very high attenuation (eg, 20-60 db) in this frequency range. A second fourth order elliptic notch filter configured as a high notch filter should attenuate the servo pattern frequency of about 5 MHz and its third harmonic of about 15 MHz. These two frequencies are the main ones. A fourth order elliptic notch filter can provide very high attenuation in this frequency range. These frequencies are unchanged because the spindle speed of a typical data storage system is accurate.
[0086]
As described above, both the magnetic and thermal components of the readback signal contain information regarding the surface characteristics of the disk surface from which the readback signal was read. A comparison of magnetic and thermal spacing signal responses for various types of disk surface defects and permanently written servo sectors is shown in Table 1 below. Separating the composite readback signal into independent magnetic and thermal signal components provides an opportunity to detect the same “unknown phenomenon” or surface defects using two independent simultaneous responses of the MR head. The use of two independent thermal and magnetic signals during disk surface analysis can greatly increase defect detection resolution and reliability. Defect detection enhancements are realized by using a two-dimensional (2D) detection technique rather than a one-dimensional (1D) technique. It is said that the one-dimensional detection method is a method using either one of the magnetic signal and the heat signal, not one of them. Using two independent magnetic and thermal signals derived from the same MR element at the same moment becomes a powerful tool for detecting and classifying unknown disk surface defects.
[0087]
Table 1 below outlines the differences between the magnetic and thermal MR head / disk spacing response for the simple disk surface defect and permanent recording shown in FIG. It should be noted that many disk surface defects, such as scratches and strikes, are usually a complex combination of several simple defects, resulting in a more complex MR head response.
[0088]
[Table 1]

[0089]
Referring to Table 1 above, the method for performing disk surface defect analysis can be performed as follows. First, a thermal scan is performed on each surface of each disk provided in the data storage system. The resulting thermal response is monitored for thermal voltages that exceed predetermined positive and negative thresholds. Since the thermal response is insensitive to magnetic air gaps and pre-written servo sectors, this process will remove the servo sector magnetic patterns and magnetic air gaps as effective surface defects. It should be noted, however, that defect analysis processes that exclusively use magnetic signals may falsely indicate the presence of surface defects when detecting magnetic gaps or permanent records at the wrong location. Activation of the thermal threshold detector during thermal operation can be attributed to three basic types of surface defects: pits, bumps, thermal bumps (TA), or a combination of these three defect types. is there.
[0090]
Next, using the magnetic response characteristics shown in Table 1 above, magnetic information is written to the disk surface location where the thermal threshold was activated to perform the magnetic defect verification procedure, such as by using an HRF or RSM method. be able to. Since only bumps and bumps due to heat should indicate a valid failure condition, the thermal detector and HRF / RSM detector should be activated simultaneously before removing the data storage system or disk that contains the defective disk inside. It must be made.
[0091]
Therefore, the use of MR element heating and cooling in response to changes in head / disk spacing does not cause mechanical damage to the disk 24 that should be sufficient to eliminate the disk and not enough to eliminate the disk. It can detect and distinguish from other non-fatal disk defects such as supposed pits and magnetic gaps. Such disk surface defect analysis can be performed in the field or in a data storage system that is fully operational prior to shipment. Further, this analysis can be performed on-site a predetermined number of times during the field life of the data storage system to perform predictive failure analysis on the data storage system. Alternatively, a disk surface defect analysis can be performed on a disk blank without a magnetic coating to avoid further processing of the defective disk blank.
[0092]
A method for classifying disk defects can be expressed by an equation using the negative peak and the positive peak of the thermal signal extracted from the MR head. One example of a defect classification method is derived from the determination that in the case of a disk depression (eg, pit), the amplitude of the thermal signal increases as the MR element heats due to increased head / disk spacing. With this class of disk defects, cooling rarely occurs and a negative polarity thermal signal should be generated by the MR head.
[0093]
In the case of approaching head / disk contact conditions, the MR element cools. As a result of the cooling of the MR element, the negative thermal voltage signal V_THOccurs. The criteria for testing a disk for mechanical damage is given by
(V +) + | V− |> T [8]
Where V + is V_THV- is V-_TH, And T is a thermal voltage threshold above which it indicates the presence of disk mechanical damage.
[0094]
The inventors have determined that the test criteria of equation [8] can be used to accurately identify the presence of existing or imminent mechanical disk damage. The application of equation [8] is reasonable when determining the presence of disk mechanical damage. This is because such damage is associated with both heating caused by disk defects that cause the head to displace and cooling caused by the MR element because it is close to the disk surface. In the case of other disk defects that are not related to upward disk surface protrusion, such as plating pits, the thermal voltage signal does not show a negative peak that is large enough to warrant concern. This is because a sufficient amount of MR element cooling has not occurred. In addition, disk defects that significantly affect magnetic signals, such as magnetic air gaps, do not lead to an appropriate thermal response. It has been common practice for magnetic data storage disk manufacturers to eliminate and discard suspicious disks that exhibit non-destructive magnetic air gaps. This is because conventional screening procedures cannot reliably verify the presence of mechanical air damage or other related disk mechanical damage.
[0095]
In general, curves can be used to distinguish between screening acceptance criteria and rejection criteria. An example of such a screening curve is given by

_- ⁿ+ C₁V₊ ^m= C₂                                           [9]
Where V_-Is the minimum thermal voltage, V₊Is the maximum thermal voltage, n, m, C₁, C₂Is a constant. For example, n = m = 2 and C₁= 1, the pass / fail curve has a radius of
[Expression 7]
It becomes a part of the circle corresponding to.
[0096]
The identification of a particular type of surface defect can be established by using a disk surface defect detection circuit. In the embodiment shown in FIG. 31, the defect identification circuit 91 is realized using an analog circuit, but it should be noted that the defect identification circuit 91 can be realized as a digital circuit or by digital signal processing. For example, as shown in the following equation, the defect detection / identification circuit 91 may have a negative thermal cooling peak V_THPositive peak V of bumpy “heating” spikes due to (cool) and heat_THTotal thermal voltage signal difference ΔV from (warm)_THBumps due to heat (TA) are detected by measuring (TA).
ΔV_TH(TA) = V_TH(Warm) -V_TH(Cool) [10]
[0097]
Thermal voltage signal V_THIs extracted from the readback signal by the signal separation / restoration module 76 and communicated to the positive peak holding circuit 71 and the negative peak holding circuit 73. The positive peak holding circuit 71 has a thermal signal V_TH(Warm) positive peak voltage is buffered, and the negative peak holding circuit 73 generates a thermal signal V_THBuffer (cool) negative peak voltage. A pit detector 75 implemented using an operational amplifier in the comparator configuration has an appropriate input threshold voltage T_(PIT)And positive peak voltage V_THCalibrated to detect surface pits by comparing with (warm).
[0098]
The hump detector 79 is similarly implemented and has an appropriate input threshold voltage T_(BUMP)And negative peak voltage V_THCalibrated to detect surface humps by comparing with (cool). Also, a bump detector 77 due to heat is realized in the same manner, and an appropriate input threshold voltage T_(TA)And the heat difference signal (V_TH(Warm) -V_TH(Cool)) is calibrated to detect thermal bumps. In an embodiment in which the positive peak holding circuit 71 and the negative peak holding circuit 73 are not used, the thermal response voltage V against a preset threshold value for defect identification._THThe maximum positive peak value and the minimum negative peak value may be continuously monitored.
[0099]
The logic levels of the three comparators 75, 77, 79 form a 3-bit word {TA, BUMP, PIT} for use in conjunction with a pass / fail decision table as shown in Table 2 below. Can be configured. As described above, if there is a disk failure, a verification procedure can be performed later by writing a magnetic track to the disk and performing a verification test of HRF and / or RSM.
[0100]
[Table 2]

[0101]
Referring to FIG. 32, there are shown various steps for performing an error recovery routine that uses the thermal signal component of the readback signal. The error recovery routine shown in flowchart form in FIG. 32 is generally appropriate for data recovery due to severe error conditions, and is usually performed after several standard error recovery routines have been executed. In step 530, defective sectors containing missing information or unreadable information are identified. In step 532, the readback signal is sampled for defective sectors or disk areas. The sampled readback signal is stored at step 534. The stored readback signal corresponding to the defective sector is RS_(DEFECT)= M_(DEFECT)+ T_(DEFECT)As shown. However, RS_(DEFECT)Represents the total readback signal obtained from the defective sector, M_(DEFECT)Represents the magnetic signal component of the read-back signal obtained from the defective sector, and T_(DEFECT)Is the read-back signal RS obtained from the defective sector_(DEFECT)Represents the thermal signal component.
[0102]
In step 536, defective sectors in the disk surface area are erased. After the erasure of the defective disk surface position is completed, the read-back signal is sampled for the erased defective sector as shown in step 538. The readback signal sample from the erased defective sector is RS_(ERASE)= T_(ERASE)Represented as: However, RS_(ERASE)Represents the total readback signal sampled from the erased defective sector and T_(ERASE)Is the read-back signal RS obtained from the erased defective sector_(ERASE)Represents the thermal signal component. However, since the erase process removes almost all of the magnetic signal component from the defective sector, the read-back signal RS_(ERASE)Note that the magnetic signal component of is not included.
[0103]
However, it should be noted that microcracks on the disk surface can maintain a small amount of magnetic flux despite the erase procedure. Thus, as shown in step 540, the thermal signal T from the read-back signal obtained from the erased defective sector._(ERASE)It may be desirable to extract and store. T_(DEFECT)Is verified by the thermal signal waveform described above in connection with FIGS._(ERASE)Note that it is almost equivalent to In step 542, the readback signal obtained for the defective sector erased in step 538 is subtracted from the readback signal obtained from the defective sector in step 532. As shown in step 542, this subtraction results in a recovered magnetic signal component M of the defective sector._(RECOVERED)Is obtained. Next, the recovery magnetic signal M_(RECOVERED)Is stored in memory or elsewhere on the data storage disk. In the final step 546, the recovered magnetic signal M_(RECOVERED)Passes from the memory through a normal data recording channel where the information is decoded into binary words.
[0104]
Of course, it should be noted that various changes and additions can be made to the above-described embodiments without departing from the scope or spirit of the invention. Accordingly, the scope of the invention should not be limited to the specific embodiments described above, but should be defined only by the complete and fair scope of the claims set forth at the beginning.
[0105]
In summary, the following matters are disclosed regarding the configuration of the present invention.
[0106]
(1) A method for processing a signal obtained from a storage medium using a magnetoresistive (MR) element proximate to the storage medium, the method comprising:
Reading a signal from a storage medium using an MR element;
Extracting from the signal a thermal signal component representative of the thermal response of the MR element.
(2) The method according to (1), wherein the thermal signal component represents a distance between the MR element and the storage medium.
(3) The method according to (2), wherein the thermal signal component varies linearly in response to a variation in the distance between the MR element and the storage medium.
(4) the signal includes a magnetic signal component;
The above (1) is characterized in that the thermal signal component is calibrated using a magnetic signal component so as to estimate a variation in a distance between the MR element and the storage medium using the thermal signal component. The method described.
(5) The signal includes a magnetic signal component, and the method includes:
Using the thermal signal component and the magnetic signal component to generate a thermal interval signal, the thermal interval signal varying in proportion to variations in the interval between the MR element and the storage medium, The method according to (1).
(6) The method according to (5) above, wherein the thermal interval signal fluctuates linearly according to fluctuations in the interval between the MR element and the storage medium.
(7) The method according to (1), wherein the thermal signal component represents a characteristic of a storage medium.
(8) The method according to (7), wherein the characteristic of the storage medium is a surface contour of the storage medium.
(9) The method according to (7), wherein the characteristic of the storage medium is an emissivity of the storage medium.
(10) The method according to (1) above, wherein the extracting step includes a step of filtering a signal in order to remove a thermal component from the signal.
(11) The method according to (10), wherein the extracting step includes the step of filtering the signal using a finite impulse response (FIR) filter.
(12) The filtering step of (11), wherein the filtering step comprises programming the FIR filter using a set of tap weights stored in a memory coupled to the FIR filter. Method.
(13) The method according to (12), wherein the filtering step includes a step of applying a window function to the set of tap weights when programming an FIR filter.
(14) The filtering step comprises:
Programming an FIR filter using a first set of tap weights to read data stored on the storage medium;
Programming the FIR filter using a second set of tap weights to read servo information stored on the storage medium.
(15) The method according to (1), wherein the signal includes servo information, and the extracted thermal signal component represents the servo information.
(16) The method according to (1), wherein the signal includes a magnetic signal component representing servo information.
(17) The method according to (1), further including a step of extracting a magnetic component from the signal.
(18) The signal includes a thermal signal component and a magnetic signal component, and the method includes:
The method of (1) above, further comprising extracting a magnetic signal component from the signal by subtracting the extracted thermal signal component from the signal.
(19) reading a signal from a defective portion of the storage medium, wherein the signal is a composite signal including a magnetic signal component and a thermal signal component;
Erasing a magnetic signal component from a defective portion of the storage medium;
Extracting a thermal signal component from the erased defective portion of the storage medium;
Subtracting the thermal signal component extracted from the composite signal to generate a reconstructed magnetic signal that substantially represents the magnetic signal component of the composite signal, the method of (1) above.
(20) In a signal separation device for an information storage device including an information storage medium,
A transducer including a magnetoresistive (MR) element;
A read channel connected to the transducer for reading a signal from the storage medium by an MR element proximate to the storage medium;
An apparatus coupled to the read channel and for extracting a thermal signal component of the signal, wherein the thermal signal component represents a thermal response of the MR element.
(21) The device according to (20), wherein the filter is a finite impulse response (FIR) filter.
(22) The apparatus according to (20) above, wherein the signal read from the storage medium includes a magnetic signal component.
(23) The apparatus according to (20), wherein the signal read from the storage medium is a magnetic signal including a thermal signal component.
(24) The apparatus according to (20) above, wherein the signal read from the storage medium is a servo information signal.
(25) The apparatus according to (20) above, wherein the extracted thermal signal component is a servo information signal.
(26) The apparatus according to (20), further comprising a magnetic signal filter coupled to the read channel and for extracting a magnetic signal component of a signal read from the storage medium.
(27) including a signal summing device coupled to the read channel and filter;
A signal adder receives the extracted thermal signal component and a composite signal read from the storage medium including the thermal signal component and the magnetic signal component, and uses the extracted signal component to generate heat from the composite signal. The apparatus according to (20), wherein the signal component is subtracted to generate a restored magnetic signal that substantially represents the magnetic signal component of the composite signal.
(28) In the information storage device,
A transducer including a magnetoresistive (MR) element;
A storage medium;
Means for moving at least one of the transducer and the storage medium to effect relative movement between the transducer and the medium, the transducer being arranged relative to the medium such that a gap separates the MR element from the medium And the means being
A read channel connected to the transducer to read signals from the medium using the MR element;
An apparatus coupled to the read channel and for extracting a thermal signal component of the signal, wherein the thermal signal component represents a thermal response of the MR element.
(29) The apparatus according to (28), wherein the filter includes a finite impulse response (FIR) filter.
(30) The apparatus of (29) above, comprising a memory coupled to the FIR filter, wherein the FIR filter is programmable using a set of tap weights stored in the memory.
(31) the memory stores a set of window parameters;
The apparatus of (29) above, wherein the FIR filter is programmed using a set of tap weights and window parameters stored in the memory.
(32) a first set of tap weights is transferred from the memory to the FIR filter when the MR element moves proximate to the data storage portion of the medium;
The apparatus of (29) above, wherein a second set of tap weights is transferred from the memory to the FIR filter when the MR element moves proximate to the servo information storage portion of the medium.
(33) The apparatus according to (28), wherein the signal read from the medium includes a magnetic signal component.
(34) The apparatus according to (28), wherein the signal read from the medium is a magnetic signal including a thermal signal component.
(35) The apparatus according to (28), wherein the signal read from the medium is a servo information signal.
(36) The apparatus according to (28), wherein the extracted thermal signal component is a servo information signal.
(37) The apparatus according to (28), further comprising a magnetic signal filter coupled to the read channel and for extracting a magnetic signal component of a signal read from the medium.
(38) including a signal summing device coupled to the read channel and filter;
A signal adder receives the extracted thermal signal component and a composite signal read from the medium including the thermal signal component and the magnetic signal component, and uses the extracted signal component to generate a thermal signal from the composite signal. The apparatus according to (28), wherein the component is subtracted to generate a restored magnetic signal that substantially represents the magnetic signal component of the composite signal.
(39) including a signal adding device and a writing element;
The MR element reads a signal from a defective portion of the storage medium, and the signal is a composite signal including a magnetic signal component and a thermal signal component;
A writing element erases a magnetic signal component from a defective portion of the storage medium;
The filter extracts a thermal signal component from the erased defective portion of the storage medium;
The apparatus according to (28), wherein the signal adder subtracts the thermal signal component extracted from the composite signal to generate a restored magnetic signal.
(40) The apparatus according to (39) above, wherein a restored magnetic signal is decoded in the read channel.
(41) In a method for inspecting a medium using a signal induced by a magnetoresistive (MR) element spaced from the medium,
Reading signals using the MR element as the MR element moves relative to the medium;
Detecting fluctuations in the thermal signal component of the signal read by the MR element;
Determining changes in the properties of the medium from variations in the thermal signal component.
(42) The method according to (41), wherein the change in the characteristics of the medium is a change in the surface contour of the medium.
(43) The method according to (41) above, wherein the change in characteristics of the medium is a change in emissivity of the medium.
(44) The method according to (41), wherein the detecting step includes a step of detecting a change in an interval between the MR element and the medium.
(45) determining intervals at a plurality of positions on a portion of the medium;
Mapping the surface relief representation of a portion of the medium using intervals determined at a plurality of locations. (46) detecting thermal signal component variations at a plurality of locations on a portion of the medium;
Mapping the representation of the portion of the media using variations in thermal signal components at a plurality of locations.
(47) the medium characteristic is a surface contour of the medium, and the method includes generating a thermal spacing signal from the thermal signal component, wherein the thermal spacing signal is a measure of the spacing between the element and the surface of the medium. The method according to (44) above, characterized by:
(48) The characteristic of the medium is a surface contour of the medium, and the method includes:
Detecting a variation in a magnetic signal component of a signal read from the medium using the MR element;
Generating a magnetic spacing signal from the magnetic signal component, wherein the magnetic spacing signal represents a measure of the spacing between the MR element and the surface of the medium;
Generating a thermal spacing signal from the thermal signal component, wherein the thermal spacing signal is calibrated using a magnetic spacing signal to represent a measure of the spacing between the MR element and the surface of the medium. The method according to (44) above.
(49) The characteristic of the medium is a surface contour of the medium, and the method includes:
Determining the frequency of the thermal signal component;
Associating the frequency of the thermal signal component with one of the surface contour features of the medium.
(50) the method includes detecting variations in a magnetic signal component of a signal read from the medium using the MR element;
The characteristic of the medium is the surface contour of the medium;
The method according to (44), wherein the determining step includes a step of determining a change in a surface contour of the medium using a variation in a thermal signal component and a variation in a magnetic signal component.
(51) In the above (44), wherein the characteristic of the medium is a surface contour of the medium, and the method includes detecting a feature of the surface contour using a variation of a thermal signal component The method described.
(52) writing a magnetic signal at a position on the medium corresponding to the position of the feature;
Reading magnetic signals using the MR element;
Detecting a feature of a surface contour using a magnetic signal read using the MR element. (53) The characteristic of the medium is a surface contour of the medium including a depression provided on the surface of the medium, and the method includes:
The method of (41) above, comprising detecting a depression provided on the surface of the medium using variation of the thermal signal component.
(54) The method according to (53) above, wherein the recess includes a groove provided on a surface of the medium.
(55) The method according to (53), wherein the recess includes a pit provided on a surface of the medium.
(56) In the information storage device,
A transducer including a magnetoresistive (MR) element;
A storage medium;
Means for moving at least one of the transducer and the storage medium to effect relative movement between the transducer and the medium, the transducer being disposed relative to the medium such that the spacing separates the MR element from the medium And the means being
A read channel connected to the transducer to read signals from the medium using the MR element;
A filter coupled to the read channel for passing the thermal signal component of the signal, wherein the thermal signal component represents the thermal response of the MR element;
And a detector coupled to the filter for detecting variations in the thermal signal component corresponding to variations in the surface contour of the storage medium.
(57) The apparatus according to (56), wherein the fluctuation of the surface contour represents servo information.
(58) The apparatus according to (56), wherein the fluctuation of the thermal signal detected by the detector corresponds to a change in the dimension of the distance between the MR element and the surface contour of the storage medium. .
(59) The above (58), wherein the detector detects a variation in a spacing dimension at a plurality of positions on a portion of the storage medium, and characterizes a variation in a surface contour of the portion of the storage medium. The device described in 1.
(60) The above (56), wherein the detector detects a variation in a thermal signal component at a plurality of positions on a portion of the storage medium, and characterizes a variation in a surface contour of the portion of the storage medium. ) Device.
(61) including an average filter coupled to the filter for converting a thermal signal component into a thermal interval signal, wherein the thermal interval signal corresponds to a dimension of an interval between the MR element and the surface of the medium. The device according to (56) above.
(62) a magnetic signal filter coupled to the filter and log filter for converting a magnetic signal component of a signal read from the storage medium using the MR element into a magnetic interval signal; Device according to (56) above, characterized in that the signal corresponds to the dimension of the spacing between the MR element and the surface of the medium.
(63) including a magnetic signal filter coupled to the filter and the log filter, for converting a magnetic signal component of a signal read from the storage medium using the MR element into a magnetic interval signal; Device according to (62), characterized in that the signal and the thermal spacing signal substantially correspond to the dimension of the spacing between the MR element and the surface of the medium.
(64) The apparatus according to (56), further including a defect association circuit for associating a frequency of the thermal signal component with one of the features of the surface contour.
(65) the detector detects a variation in a magnetic signal component of a signal read from the storage medium using the MR element, corresponding to a surface contour feature of the storage medium;
The apparatus according to (56), wherein the detector detects a variation in a thermal signal component corresponding to a feature of a surface contour.
(66) The apparatus according to (56), wherein the detector detects a feature of a surface contour of the storage medium as a variation of a thermal signal component.
(67) In a method of reading a signal from a medium using a single magnetoresistive (MR) element of a transducer, a first type of information and a second type of information contained in the signal from the medium using the MR element Reading the information simultaneously, wherein the first type of information is represented as a magnetic component of the signal and the second type of information is represented as a thermal component of the signal.
(68) The method according to (67), wherein the first type of information includes data, and the second type of information includes transducer position information.
[Brief description of the drawings]
FIG. 1 is a block diagram of an apparatus for extracting a thermal signal and a magnetic signal from a readback signal induced by an MR head.
FIG. 2 is a plan view of the data storage system with its upper housing cover removed.
FIG. 3 is a flow diagram illustrating a method for obtaining and using a thermal signal component of a readback signal.
FIG. C. It is a figure which shows the readback signal induced | guided | derived with MR head which shows the base line.
FIG. 5 shows the restored D.C. after processing by the signal separation / modulation module. C. FIG. 5 is a diagram showing the readback signal of FIG. 4 showing a base line.
FIG. 6 is a diagram showing a thermal signal extracted from a read-back signal induced by an MR element at a specific track position.
FIG. 7 is a diagram showing a read-back signal obtained from the same track position after AC erasure of magnetic information.
FIG. 8 is an enlarged side view of a data storage disk showing various surface defects and features and MR element thermal and magnetic spacing responses to such defects and features.
FIG. 9 illustrates a readback signal indicating a head / disk contact event.
FIG. 10 extracts a thermal signal and a magnetic signal from a read-back signal induced by an MR element, and outputs a D.D. C. FIG. 2 is a block diagram of a signal separation / modulation module for restoring a baseline.
FIG. 11 shows the D. of the magnetic read-back signal. C. FIG. 2 is a block diagram of a signal separation / modulation module for restoring a baseline.
FIG. 12 is a block diagram of a system for selectively communicating a readback signal to a signal separation / modulation module.
FIG. 13 is a diagram showing an absolute value response of a finite impulse response (FIR) filter used in the signal separation / restoration module.
FIG. 14 shows the phase response of a finite impulse response (FIR) filter used in the signal separation / restoration module.
FIG. 15 is a diagram showing a read-back signal induced by the MR head.
FIG. 16 is a diagram showing a restored magnetic signal component of a read-back signal.
FIG. 17 is a diagram showing an unrestored magnetic signal component of a read-back signal.
FIG. 18 is a diagram showing an absolute value response of a windowed FIR filter used in the signal separation / restoration module.
FIG. 19 is a diagram showing the phase response of a windowed FIR filter used in the signal separation / restoration module.
FIG. 20 shows the absolute value response of the high pass filter behavior of a typical AE module.
FIG. 21 is a diagram showing the phase response of the high-pass filter behavior of a typical AE module.
FIG. 22 shows a comparison of absolute response of high pass filter behavior between a typical AE module and an inverse filter having a transfer function opposite to that of the effective high pass filter of the AE module.
FIG. 23 shows a comparison of the phase response of the high pass filter behavior of a typical AE module and an inverse filter having a transfer function opposite to that of the effective high pass filter of the AE module.
24 is a signal flow diagram representing the inverse filter of FIGS. 23 and 24. FIG.
FIG. 25 is a diagram showing a waveform due to a disk surface pit generated at a processing point in the signal separation / restoration module;
FIG. 26 is a diagram showing a waveform due to a disk surface pit generated at a processing point in the signal separation / restoration module.
FIG. 27 is a diagram showing a waveform due to a disk surface pit generated at a processing point in the signal separation / restoration module;
FIG. 28 is a diagram showing a close correspondence between a magnetic spacing signal and a thermal head / disk spacing signal associated with a disk surface defect.
FIG. 29 is a block diagram of another embodiment of a signal separation / restoration module that uses an infinite impulse response (IIR) filter.
FIG. 30 illustrates a magnetic spacing signal and a thermal spacing signal associated with a head / disk contact event.
FIG. 31 is a block diagram of a defect classification circuit.
FIG. 32 is a flow diagram of an error recovery process that uses a thermal signal for a readback signal.
[Explanation of symbols]
24 data storage disk
60 Readback signal
61a Magnetic signal component
63 Heat signal
70 devices
72 Magnetoresistive (MR) element
74 AE module
76 Signal separation / restoration module
78 Magnetic signal
80 Heat signal

Claims (24)

  A magnetic disk;
  A magnetoresistive head for reading information from the magnetic disk;
  An amplifier for amplifying the read signal of the magnetoresistive head;
  A high-pass filter that passes high-frequency components of the output of the amplifier;
  An analog-to-digital converter that digitizes the output of the high-pass filter;
  A programmable filter for restoring a magnetic signal by passing a high-frequency component from the output of the analog-digital converter;
  An amplitude detector for detecting the peak-to-peak amplitude of the output of the analog-digital converter;
  A log device that takes the logarithm of the output of the amplitude detector and generates a magnetic spacing signal linearly related to the spacing between the magnetoresistive head and the magnetic disk;
  An inverse filter having a transfer function opposite to that of the high-pass filter, which receives the output of the analog-digital converter;
  A thermal signal extraction filter for extracting a thermal signal from the output of the inverse filter;
  An average filter that inputs a thermal signal from the thermal signal extraction filter and generates a thermal interval signal linearly related to the interval between the magnetoresistive head and the magnetic disk;
A magnetic disk storage device comprising:   The magnetic disk storage device of claim 1, wherein the programmable filter is a high-pass filter.   2. The magnetic disk storage device according to claim 1, wherein the thermal signal extraction filter is a low-pass filter.   The magnetic disk storage device according to claim 1, wherein the thermal signal extraction filter is programmable and is a finite impulse response filter.   2. The magnetic disk storage device according to claim 1, wherein the inverse filter and the thermal signal extraction filter are constituted by a single infinite impulse response filter.   2. The magnetic disk storage device according to claim 1, wherein the average filter is a digital operation smoothing averaging filter.   2. A magnetic disk storage device according to claim 1, further comprising means for calibrating the thermal interval signal using the magnetic interval signal in order to accurately reflect the interval between the magnetoresistive head and the magnetic disk. apparatus.   The magnetic disk has a plurality of recording tracks, the recording tracks have a plurality of servo sectors and data sectors, and the magnetic interval signal and the thermal interval signal are generated when the servo sector is read. Item 2. A magnetic disk storage device according to Item 1.   A magnetic disk;
  A magnetoresistive head for reading information from the magnetic disk;
  An amplifier for amplifying the read signal of the magnetoresistive head;
  A high pass filter for removing low frequency components from the output of the amplifier;
  An analog-to-digital converter that digitizes the output of the high-pass filter;
  A programmable filter for restoring a magnetic signal by passing a high-frequency component from the output of the analog-digital converter;
  An amplitude detector for detecting the peak-to-peak amplitude of the output of the analog-digital converter;
  A log device that takes the logarithm of the output of the amplitude detector and generates a magnetic spacing signal linearly related to the spacing between the magnetoresistive head and the magnetic disk;
  A thermal signal extraction filter for extracting a thermal signal from the output of the amplifier;
  An average filter that inputs a thermal signal from the thermal signal extraction filter and generates a thermal interval signal linearly related to the interval between the magnetoresistive head and the magnetic disk;
A magnetic disk storage device comprising:   The magnetic disk storage device of claim 9, wherein the programmable filter is a high-pass filter.   The magnetic disk storage device according to claim 9, wherein the thermal signal extraction filter is a low-pass filter.   10. The magnetic disk storage device according to claim 9, wherein the thermal signal extraction filter is programmable and is a finite impulse response filter.   10. The magnetic disk storage device according to claim 9, wherein the inverse filter and the thermal signal extraction filter are constituted by a single infinite impulse response filter.   10. The magnetic disk storage device according to claim 9, wherein the average filter is a digital operation smooth averaging filter.   10. The magnetic disk storage of claim 9, further comprising means for calibrating the thermal interval signal using the magnetic interval signal to accurately reflect the interval between the magnetoresistive head and the magnetic disk. apparatus.   The magnetic disk has a plurality of recording tracks, the recording tracks have a plurality of servo sectors and data sectors, and the magnetic interval signal and the thermal interval signal are generated when the servo sector is read. Item 10. A magnetic disk storage device according to Item 9.   A magnetic disk;
  A magnetoresistive head for reading information from the magnetic disk;
  An amplifier for amplifying the read signal of the magnetoresistive head;
  A high-pass filter that passes high-frequency components of the output of the amplifier;
  An analog-to-digital converter that digitizes the output of the high-pass filter;
  A programmable filter for restoring a magnetic signal by passing a high-frequency component from the output of the analog-digital converter;
  An amplitude detector for detecting the peak-to-peak amplitude of the output of the analog-digital converter;
  A log device that takes the logarithm of the output of the amplitude detector and generates a magnetic spacing signal linearly related to the spacing between the magnetoresistive head and the magnetic disk;
  An inverse filter having a transfer function opposite to that of the high-pass filter, which receives the output of the analog-digital converter;
  A thermal signal extraction filter for extracting a thermal signal from the output of the inverse filter;
  An average filter that inputs a thermal signal from the thermal signal extraction filter and generates a thermal interval signal linearly related to the interval between the magnetoresistive head and the magnetic disk;
  A positive peak holding circuit for holding a positive peak voltage of the thermal signal from the thermal signal extraction filter;
  A negative peak holding circuit for holding a negative peak voltage of the thermal signal from the thermal signal extraction filter;
  The holding voltage of the positive peak holding circuit and the first threshold value are compared, and the magnetic disk A pit detector that detects pits on the surface;
  A detector for comparing a holding voltage of the negative peak holding circuit with a second threshold value to detect a hump on the surface of the magnetic disk;
  A difference detector between the positive peak voltage and the negative peak voltage and a third threshold value to detect a bump due to heat on the surface of the magnetic disk;
A magnetic disk storage device comprising:   The magnetic disk storage device of claim 17, wherein the programmable filter is a high pass filter.   18. The magnetic disk storage device according to claim 17, wherein the thermal signal extraction filter is a low-pass filter.   18. The magnetic disk storage device according to claim 17, wherein the thermal signal extraction filter is programmable and is a finite impulse response filter.   18. The magnetic disk storage device according to claim 17, wherein the inverse filter and the thermal signal extraction filter are constituted by a single infinite impulse response filter.   18. The magnetic disk storage device according to claim 17, wherein the average filter is a digital operation smoothing averaging filter.   18. The magnetic disk storage device of claim 17, further comprising means for calibrating the thermal interval signal using the magnetic interval signal to accurately reflect the interval between the magnetoresistive head and the magnetic disk. apparatus.   The magnetic disk has a plurality of recording tracks, the recording tracks have a plurality of servo sectors and data sectors, and the magnetic interval signal and the thermal interval signal are generated when the servo sector is read. Item 18. A magnetic disk storage device according to Item 17.

Images