Search Images Maps Play Gmail Drive Calendar Translate More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberCN101361125 B
Publication typeGrant
Application numberCN 200780001748
PCT numberPCT/EP2007/050212
Publication date12 Jan 2011
Filing date10 Jan 2007
Priority date26 Jan 2006
Also published asCA2636848A1, CA2636848C, CN101361125A, DE602007001872D1, EP1984919A1, EP1984919B1, US7245450, US20070171565, WO2007085529A1
Publication number200780001748.1, CN 101361125 B, CN 101361125B, CN 200780001748, CN-B-101361125, CN101361125 B, CN101361125B, CN200780001748, CN200780001748.1, PCT/2007/50212, PCT/EP/2007/050212, PCT/EP/2007/50212, PCT/EP/7/050212, PCT/EP/7/50212, PCT/EP2007/050212, PCT/EP2007/50212, PCT/EP2007050212, PCT/EP200750212, PCT/EP7/050212, PCT/EP7/50212, PCT/EP7050212, PCT/EP750212
InventorsE埃莱夫特里乌, G凯鲁比尼, J耶利托, R哈钦斯
Applicant国际商业机器公司
Export CitationBiBTeX, EndNote, RefMan
External Links: SIPO, Espacenet
Synchronous servo channel for longitudinal position detection and position error signal generation in tape drive systems
CN 101361125 B
Abstract
A fully synchronous longitudinal position (LPOS) detection system is provided for improving the reliability of servo channels in tape systems. The system is based on the interpolation of the servo channel output signal, which is sampled by an analog-to-digital converter (ADC) at a fixed sampling rate, using a clock at a nominal frequency, so that interpolated signal samples are obtained at a predetermined fixed rate, independent of tape velocity. This predetermined fixed rate is defined in terms of samples per unit of length, as opposed to samples per unit of time, which is the measure of theADC sampling rate. The resolution with which the servo channel signal is obtained at the interpolator output is thus determined by the step interpolation distance.
Claims(12)  translated from Chinese
  1. 一种在磁带存储系统中进行纵向位置(LPOS)探测的方法,包括:对磁带存储系统进行操作,由此使磁带以一定速度纵向通过伺服读磁头,所述磁带包括记录下来的嵌有LPOS信息的伺服脉冲图形;确定预定的固定采样率,该采样率按照通过所述伺服读磁头的单位磁带长度上的样本来定义;从所述伺服读磁头产生伺服信道输出信号;产生具有标称频率的时钟信号;以所述标称时钟频率采用模拟‑数字转换器(ADC)对所述伺服信道输出信号进行采样;对来自所述ADC的伺服信道输出信号样本进行插值;以不依赖于所述磁带速度的所述固定采样率产生插值信号样本序列;以及从所述插值信号样本序列中产生LPOS信号。 A longitudinal position (LPOS) detection method in a tape storage system, comprising: a tape storage system, whereby the tape lengthwise at a constant speed by a servo read head, the tape comprising recorded information embedded with LPOS servo pulse pattern; determining a predetermined fixed sampling rate, the sampling rate is defined according to the sample by the servo read head of the tape unit length; generating a servo channel output signal from the servo read head; generating a nominal frequency clock signal; said nominal clock frequency to analog - digital converter (ADC) for the servo channel output signal samples; the servo channel output signal samples from the ADC interpolating; to not depend on the magnetic tape The rate of fixed sampling rate to generate the interpolated signal sample sequence; and generating LPOS signal from the interpolated signal sample sequence.
  2. 2.根据权利要求1所述的方法,其中,产生LPOS信号包括: 获得基于时序的伺服脉冲;以及测量各个双位的过零点之间的距离以产生位置误差信号。 2. The method of claim 1, wherein the generating LPOS signal comprises: obtaining timing-based servo bursts; and each dibit measurement distance between the zero crossings of the position error signal to generate.
  3. 3.根据权利要求2所述的方法,其中,所述LPOS信号在磁带速度变化和恒定期间产生。 3. The method according to claim 2, wherein the LPOS signal and the constant change in the tape speed during production.
  4. 4.根据权利要求3所述的方法,其中,用于LPOS探测的近似最大速度由vmax = Lp ^ (fs/2)来表示,其中,Lp是所述伺服脉冲图形中的磁性跃变之间的最小距离,单位为微米, 而fs为所述固定采样率,单位为MHz。 4. The method of claim 3, wherein an approximate maximum velocity for LPOS detection by vmax = Lp ^ (fs / 2) is represented, wherein, Lp between the servo burst patterns of magnetic transition The minimum distance, in microns, and fs is the fixed sampling rate, in units of MHz.
  5. 5. 一种用于磁性数据存储带的同步纵向位置(LPOS)探测系统,包括: 时钟,用于以标称频率产生时钟信号;用于确定预定的固定采样率的装置,该固定采样率按照纵向通过伺服读磁头的单位磁带长度上的样本来定义,所述磁带包括记录下来的嵌有LPOS信息的伺服脉冲图形; 模拟-数字转换器(ADC),包括: 用于接收所述时钟信号的输入端; 用于从所述伺服读磁头接收伺服信道输出信号的输入端; 用于以所述标称时钟频率对所述伺服信道输出信号进行采样的装置;以及用于输出伺服信号样本的装置;时序基准发生器,其具有用来接收所述时钟信号的输入端; 脉冲插值器,其对所述时序基准发生器进行响应,包括: 用来从所述ADC接收所述伺服信号样本的第一输入端;以及用来以独立于磁带速度的固定采样率对所述伺服信号样本进行插值并输出插值信号样本序列的装置;以及LPOS探测器,包括:用来从所述脉冲插值器接收所述插值信号样本序列的输入端; 用来向所述脉冲插值器的第二输入端发送控制信号的第一输出端;以及用来从所述插值信号样本序列产生LPOS输出信号的装置。 5. A synchronous longitudinal position of the magnetic data storage strip (LPOS) detection system, comprising: a clock for generating a clock signal at a nominal frequency; means for determining a predetermined fixed sampling rate used, in accordance with the fixed sampling rate longitudinal read head by servo samples on the length of the tape unit is defined, the tape comprising recorded servo burst patterns embedded with LPOS information; analog - digital converter (ADC), comprising: means for receiving said clock signal, inputs; from the servo read head for receiving the servo channel output signal input terminal; to the nominal clock frequency of the servo channel output signal means for sampling; and means for outputting the servo signal samples ; timing reference generator, having for receiving said clock signal input terminal; pulse interpolator, its response to the timing reference generator, comprising: means for receiving said first servo signal samples from the ADC an input terminal; and means independent of the tape speed fixed sampling rate of the servo signal and outputting interpolated samples interpolated signal sample sequence; and LPOS detector comprising: means for receiving from the pulse interpolator said interpolation input signal sample sequence; and a first output terminal for transmitting the control signal to the second input terminal of the pulse interpolator; and means for generating the LPOS output signal from the sequence of interpolated signal samples.
  6. 6.根据权利要求5所述的LPOS探测系统,其中,所述用来产生LPOS输出信号的装置包括用于测量各个双位的过零点之间的距离的装置。 6. A LPOS detection system of claim 5, wherein said means for generating the LPOS output signal comprises means for each dibit distance between the zero crossings to measure.
  7. 7.根据权利要求5所述的LPOS探测系统,其中,所述LPOS输出信号在磁带速度变化和恒定期间产生。 7. A LPOS detection system of claim 5, wherein the LPOS output signal and the constant change in tape speed during production.
  8. 8.根据权利要求6所述的LPOS探测系统,其中,用于LPOS探测的近似最大速度由vmax =Lp * (fs/2)表示,其中,Lp为所述伺服脉冲图形中的磁性跃变之间的最小距离,单位为微米,而fs为所述固定采样率,单位为MHz。 LPOS detection system according to claim 6, wherein, wherein the approximate maximum velocity for LPOS detection by vmax = Lp * (fs / 2), where, Lp is the servo burst pattern of magnetic transitions The minimum distance between the units of micrometers, and fs is the fixed sampling rate, in units of MHz.
  9. 9. 一种在磁带存储系统中进行纵向位置(LPOS)探测的装置,包括:用于对磁带存储系统进行操作,由此使磁带以一定速度纵向通过伺服读磁头的装置, 所述磁带包括记录下来的嵌有LPOS信息的伺服脉冲图形;用于确定预定的固定采样率的装置,该采样率按照通过所述伺服读磁头的单位磁带长度上的样本来定义;用于从所述伺服读磁头产生伺服信道输出信号的装置; 用于产生具有标称频率的时钟信号的装置;用于以所述标称时钟频率采用模拟-数字转换器(ADC)对所述伺服信道输出信号进行采样的装置;用于对来自所述ADC的所述伺服信道输出信号样本进行插值的装置; 用于以不依赖于所述磁带速度的所述固定采样率产生插值信号样本序列的装置;以及用于从所述插值信号样本序列中产生LPOS信号的装置。 A longitudinal position (LPOS) detection means, in a tape storage system comprising: a tape storage system, whereby the tape lengthwise at a constant speed by means of servo read head, the magnetic tape recording comprising down pulse pattern embedded servo LPOS information; means for determining a predetermined fixed sampling rate for the sampling rate is defined according to the sample by the servo read head of unit length of tape on; for from the servo read head means for generating a servo channel output signal; means having a nominal frequency of a clock signal for generating; to the nominal clock frequency analog to - digital converter (ADC) for the servo channel output signal sampling means ; of the servo channel output signal samples from the ADC for interpolating means; means it does not depend on the magnetic tape speed fixed sampling rate to generate the interpolated signal sample sequence for; and means from the said interpolated signal sample sequence generating means LPOS signal.
  10. 10.根据权利要求9所述的装置,其中,产生LPOS信号包括: 获得基于时序的伺服脉冲;以及测量各个双位的过零点之间的距离以产生位置误差信号。 10. The apparatus according to claim 9, wherein the generating LPOS signal comprises: obtaining timing-based servo bursts; and the distance between the zero crossings of each dibit measurement to generate a position error signal.
  11. 11.根据权利要求10所述的装置,其中,所述LPOS信号在磁带速度变化和恒定期间产生。 11. The apparatus according to claim 10, wherein the LPOS signal and the constant change in tape speed during production.
  12. 12.根据权利要求11所述的装置,其中,用于LPOS探测的近似最大速度由vmax = Lp 女(fs/2)来表示,其中,Lp是所述伺服脉冲图形中的磁性跃变之间的最小距离,单位为微米,而fs为所述固定采样率,单位为MHz。 12. The apparatus of claim 11, wherein an approximate maximum velocity for LPOS detection vmax = Lp M (fs / 2) is represented, wherein, Lp is the servo burst pattern by the magnetic transitions between The minimum distance, in microns, and fs is the fixed sampling rate, in units of MHz.
Description  translated from Chinese

磁带驱动器系统中用于探测纵向位置和产生位置误差信号 Tape drive system for detecting the longitudinal position and generating a position error signal

的同步伺服信道 The synchronous servo channel

技术领域 Technical Field

[0001] 本发明一般地说涉及到磁带存储系统中的伺服信道体系结构,具体地说涉及到一种同步伺服信道体系结构。 [0001] The present invention relates generally to a tape storage system servo channel architecture, and more particularly relates to a synchronous servo channel architecture.

背景技术 Background

[0002] 基于时序的伺服系统(timing-based servo, TBS)是90年代后期特别为线性磁带驱动器发展起来的一种技术。 [0002] The timing based servo system (timing-based servo, TBS) is a technology specially developed late for linear tape drive in the 1990s. 在TBS系统中,所记录的伺服图形由两种按不同方向倾斜的跃变(transition)构成,磁头位置根据读取伺服图形的窄磁头所产生的脉冲的相对时序来获得。 In the TBS system servopattern recorded from two different directions inclined transition (transition) constitution, the relative timing of pulses according to the read head position servo pattern generated by a narrow head to obtain. TBS图形也允许对附加的纵向位置(longitudinal position, LP0S)信息进行编码而不影响横向位置误差信号(position error signal,PES)的产生。 TBS also allows for additional graphics longitudinal position (longitudinal position, LP0S) to encode information without affecting the lateral position error signal (position error signal, PES) production. 这一点可以通过使所述跃变从其标称的图形位置发生偏移来实现,如图1所示。 This is done by making the transition from the nominal pattern position shifted to achieve, as shown in FIG. 在磁带系统中,通常有两个专用伺服信道,从这两个专用伺服信道中可以获得LPOS信息和PES。 In tape systems there are usually two dedicated servo channels from the available two dedicated servo channels LPOS information and PES. 用于线性磁带系统的基于时序的磁道跟踪伺服系统被线性磁带开放协议(Linear tape open,LT0)协会所采纳,作为所谓的LTO磁带驱动系统的标准。 For linear tape system based on the timing of the track following servo system is adopted by Linear Tape Open (Linear tape open, LT0) Association, as so-called LTO tape drive system standard.

[0003] 通常,LPOS信息的探测基于伺服信道输出中双位信号样本(dibitsignal samples)的峰位移动。 [0003] Typically, the probe LPOS information based on double servo channel output signal sample (dibitsignal samples) of peak movement. 这种方法有下面的严重局限: This approach has serious limitations the following:

[0004] a)如果希望有不依赖于磁带速度的恒定采样率(单位为每微米的样本数),那么A/D转换器的采样频率必须随磁带速度而变化。 [0004] a) If you want to have the tape speed does not depend on a constant sampling rate (in number of samples per micron), then A / D converter sampling frequency must change with the tape speed. 如果采用固定的采样频率,那么,所述采样率依赖于磁带速度。 If a fixed sampling frequency, then the sampling rate is dependent on the speed of the tape.

[0005] b)选择固定采样率的结果是,LPOS探测器(峰探测器)中每个双位响应的样本数根据磁带速度是可变的。 [0005] b) choose a fixed sample rate result is that the number of samples LPOS detector (peak detector) in each dibit response based on the tape speed is variable.

[0006] c)在加速和减速过程中,即在磁带速度朝着目标速度变化的过程中,不可能进行可靠的LPOS探测。 [0006] c) the acceleration and deceleration that the tape speed toward the target speed the process of change can not be reliably LPOS detection.

[0007] d)对于使用脉冲位置调制(pulse position modulation, PPM)技术所产生的LPOS图形,峰探测不是最适宜的探测方案。 [0007] d) For pulse position modulation (pulse position modulation, PPM) LPOS graphics generated by the peak detection is not the most appropriate detection scheme.

[0008] e)在伺服信道的输出中不可能监视信号的时间演化,因为没有时间参考。 [0008] e) at the output of the servo channel is not possible to monitor the time evolution of the signal, because there is no time reference.

[0009] f)不可能度量LPOS探测过程的可靠性。 [0009] f) can not measure the reliability of LPOS detection process.

[0010] 为了使用现有技术中的LPOS异步体系结构(如图2所示)在磁带速度很高时获得充分的分辨率,要求ADC的采样频率较高。 [0010] In order to use the prior art LPOS asynchronous architecture (shown in Figure 2) to obtain a sufficient resolution at high tape speed, it requires a higher ADC sampling frequency. 例如,如果最高目标速度为ν = 12. 5m/s,那么,设ADC的采样频率为15MHz,可以获得0.83 μ m的分辨率。 For example, if the maximum target speed of ν = 12. 5m / s, then set up the ADC sampling frequency is 15MHz, you can get 0.83 μ m resolution. 很清楚,在对+/-0. 25 μ m的LPOS脉冲位置调制进行分解时,这样的分辨率是不够的。 It is clear that at the time of +/- 0. 25 LPOS pulse position modulation μ m of decomposition, such a resolution is not enough. 具体说,使用所述异步方法,需要250MHz的采样频率来获得0. 05 μ m的分辨率。 Specifically, using the asynchronous method, you need 250MHz sampling frequency to obtain 0. 05 μ m resolution.

发明内容 DISCLOSURE

[0011] 本发明提供一种完全同步的纵向位置(LPOS)探测系统,用来提高磁带系统中的 [0011] longitudinal position (LPOS) the present invention provides a fully synchronous detection system to improve tape systems

4伺服信道的可靠性。 Reliability 4 servo channels. 本发明基于伺服信道输出信号的插值,所述伺服信道输出信号由模拟-数字转换器(ADC)使用具有标称频率的时钟信号以固定的采样率进行采样,使得插值信号样本能够不依赖于磁带速度而以预定的固定采样率获得。 The present invention is based on interpolation of the servo channel output signal, the servo channel output signal from the analog - digital converter (ADC) using a nominal frequency of a clock signal at a fixed sampling rate sampling, so that samples can be interpolated signal does not depend on the magnetic tape speed to obtain a predetermined fixed sampling rate. 这个预定的固定采样率按照单位长度上的样本数而不是按照单位时间上的样本数来定义,而按照单位时间上的样本数来定义的采样率是用来度量ADC的采样率的。 This predetermined fixed sampling rate according to the number of samples per unit length rather than by the number of samples per unit time be defined, in accordance with the number of samples per unit time on the defined sampling rate is used to measure the ADC sampling rate. 因此,在插值器输出中所获得的伺服信道信号的分辨率由标称步插值距离来决定。 Therefore, the resolution in the interpolated output in the servo channel signal obtained by the step interpolation distance to decide.

[0012] 从第一个方面来看,本发明提供一种在磁带存储系统中进行纵向位置(LPOS)探测的方法,包括:对磁带存储系统进行操作,由此使磁带以一定速度纵向通过伺服读磁头, 其中所述磁带包括记录下来的嵌有LPOS信息的伺服脉冲图形;确立预定的固定采样率,该采样率按照经过所述伺服读磁头的单位磁带长度上的样本来定义;从所述伺服读磁头产生伺服信道输出信号;产生具有标称频率的时钟信号;以所述标称时钟频率采用模拟-数字转换器(ADC)对所述伺服信道输出信号进行采样;对来自所述ADC的伺服信道输出信号样本进行插值;以不依赖于所述磁带速度的固定采样率产生插值信号样本序列;以及从所述插值信号样本序列中产生LPOS信号。 [0012] From a first aspect, the present invention provides a longitudinal position (LPOS) detection in a magnetic tape storage system, comprising: a tape storage system, whereby the tape lengthwise at a constant speed by the servo read head, wherein the tape includes recorded servo information embedded with LPOS pulse pattern; establishing a predetermined fixed sampling rate, the sampling rate through the sample in accordance with the servo read head of the tape unit length is defined; from the servo read head to generate a servo channel output signal; generating a clock signal having a nominal frequency; to the nominal clock frequency analog - digital converter (ADC) for the servo channel output signal is sampled; for from the ADC's servo channel output signal samples are interpolated; not depending on the tape speed fixed sampling rate to generate interpolated signal sample sequence; and generating the LPOS signal from the sequence of interpolated signal samples.

[0013] 本发明优选提供一种方法,其中,产生所述LPOS信号包括:获得基于时序的伺服脉冲;以及测量各个双位(dibit)的过零点之间的距离以产生位置误差信号。 [0013] Preferably the present invention provides a method wherein generating the LPOS signal comprises: obtaining timing-based servo bursts; and each dibit measurement (dibit) over the distance between the zero position to generate an error signal.

[0014] 本发明优选提供一种方法,其中,所述LPOS信号在磁带速度变化和恒定期间产生。 [0014] Preferably the present invention provides a method wherein the LPOS signal and the constant change in the tape speed during production.

[0015] 本发明优选提供一种方法,其中,用于LPOS探测的近似最大速度由vmax = Lp ^ (fs/2)来表示,其中,Lp是所述伺服脉冲图形中的磁性跃变之间的最小距离,单位为微米, 而fs为所述固定采样率,单位为MHz。 [0015] Preferably the present invention provides a method wherein an approximate maximum velocity for LPOS detection by vmax = Lp ^ (fs / 2) is represented, wherein, Lp between the servo burst patterns of magnetic transition The minimum distance, in microns, and fs is the fixed sampling rate, in units of MHz.

[0016] 从第二个方面来看,本发明提供一种用于磁性数据存储带的同步纵向位置(LPOS) 探测系统,包括:时钟,其以标称频率产生时钟信号;用于确立预定的固定采样率的装置, 其中该采样率按照纵向通过伺服读磁头的单位磁带长度上的样本来定义,所述磁带包括记录下来的嵌有LPOS信息的伺服脉冲图形;模拟_数字转换器(ADC),包括:用于接收所述时钟信号的输入端;用于从所述伺服读磁头接收伺服信道输出信号的输入端;用于以所述标称时钟频率对所述伺服信道输出信号进行采样的装置;以及用于输出伺服信号样本的装置;时序基准发生器,其具有用来接收所述时钟信号的输入端;脉冲插值器,其对所述时序基准发生器产生响应,并包括:用来从所述ADC接收所述伺服信号样本的第一输入端;以及用来独立于磁带速度以所述固定采样率对所述伺服信号样本进行插值并输出插值信号样本序列的装置;以及LPOS探测器,其包括:用来从所述脉冲插值器接收所述插值信号样本序列的输入端;用来为所述脉冲插值器的第二输入端发送控制信号的第一输出端;以及用来从所述插值信号样本序列产生LPOS输出信号的装置。 [0016] From a second aspect, the present invention provides a method for synchronizing the longitudinal position of the magnetic data storage tape (LPOS) detection system, comprising: a clock that generates a clock signal at a nominal frequency; means for establishing a predetermined means a fixed sampling rate, wherein the sampling rate according to the longitudinal reading servo samples per unit length of the tape on the head to define, the tape comprising recorded servo burst patterns embedded with LPOS information; _ analog-digital converter (ADC) comprising: means for receiving an input of the clock signal; means for receiving said servo read head from the servo channel output signal input terminal; for to the nominal clock frequency of the servo channel output signal samples means; and means for outputting servo signal samples; timing reference generator, having for receiving said clock signal input terminal; pulse interpolator that generates a response to the timing reference generator, and comprising: means received from the first input of the ADC sample servo signal; and means independent of the tape speed to the fixed sampling rate of the servo signal and outputting interpolated samples interpolated signal sample sequence; and LPOS detector comprising: means for receiving the sequence of interpolated signal samples from said pulse input terminal interpolator; the pulse used to interpolator transmitting a first input terminal of a second output of the control signal; and means from the said interpolated signal sample sequence LPOS output signal generating means.

[0017] 本发明优选提供一种LPOS探测系统,其中,所述用来产生LPOS输出信号的装置包括用于测量各个双位的过零点之间的距离的装置。 Preferably [0017] The present invention provides a LPOS detection system, wherein said means for generating the LPOS output signal comprises means for each dibit distance between the zero crossings to measure.

[0018] 本发明优选提供一种LPOS探测系统,其中,所述LPOS输出信号在磁带速度变化和恒定期间产生。 [0018] Preferably the present invention provides a LPOS detection system, wherein the LPOS output signal and the constant change in tape speed during production.

[0019] 本发明优选提供一种LPOS探测系统,其中,用于LPOS探测的近似最大速度由vmax =Lp * (fs/2)来表示,其中,Lp为所述伺服脉冲图形中的各磁性跃变之间的最小距离,单位为微米,而fs为所述固定采样率,单位为MHz。 [0019] Preferably the present invention provides a LPOS detection system wherein an approximate maximum velocity for LPOS detection by vmax = Lp * (fs / 2) is represented, wherein, Lp is the servo burst patterns of the respective magnetic Yue The minimum distance between the units becomes m, and fs is the fixed sampling rate, in units of MHz.

[0020] 从第三个方面来看,本发明提供一种可以载入数字计算机内存中的计算机程序, 包括软件代码部分,在所述产品运行于计算机上时,该软件代码部分被运行以执行上述本发明中的各个步骤。 [0020] from a third aspect, the present invention provides a digital computer memory can be loaded into a computer program, including software code portions, at the time on the computer, the software code portions are run on the product runs in execution according to the present invention, the various steps.

[0021] 从第四个方面来看,本发明提供一种可编程计算机中所使用的计算机可读介质上的计算机程序产品,所述计算机程序产品具有嵌入其中的计算机可读的代码,用于磁带存储系统中的纵向位置(LPOS)探测,所述计算机可读的代码所包含的指令可以用于:对磁带存储系统进行操作,由此使磁带以某个速度纵向通过伺服读磁头,所述磁带包括记录下来的嵌有LPOS信息的伺服脉冲图形;确立预定的固定采样率,该采样率按照通过所述伺服读磁头的单位磁带长度上的样本来定义;从所述伺服读磁头产生伺服信道输出信号;产生具有标称频率的时钟信号;以所述标称时钟频率采用模拟-数字转换器(ADC)对所述伺服信道输出信号进行采样;对来自所述ADC的伺服信道输出信号样本进行插值;以不依赖于所述磁带速度的固定采样率产生插值信号样本序列;以及从所述插值信号样本序列中产生LPOS信号。 [0021] from a fourth aspect, the present invention provides a computer-programmable computer used in a computer program product readable medium, the computer program product having embedded therein a computer readable code, for Tape storage systems longitudinal position (LPOS) detection, instructing the computer readable code that can be used include: a tape storage system, whereby the tape with a longitudinal velocity servo read head, the recorded tape including embedded servo pulse pattern LPOS information; establishing a predetermined fixed sampling rate, the sampling rate is defined according to the sample by the servo read head of the tape unit length; generating a servo channel from the servo read head output signal; generating a clock signal having a nominal frequency; to the nominal clock frequency analog - digital converter (ADC) for the servo channel output signal is sampled; the servo channel output signal samples from the ADC's interpolation; not depending upon the tape speed fixed sampling rate to generate interpolated signal sample sequence; and generating the LPOS signal from the sequence of interpolated signal samples.

[0022] 本发明优选提供一种计算机程序产品,其中,用于产生所述LPOS信号的指令包括:获得基于时序的伺服脉冲的指令;以及测量各个双位的过零点之间的距离以产生位置误差信号的指令。 [0022] Preferably the present invention provides a computer program product, wherein said instructions for generating the LPOS signal comprises: obtaining timing-based servo bursts instruction; and each dibit measurement distance between the zero crossings of the position to produce instruction error signal.

[0023] 本发明优选提供一种计算机程序产品,其中,所述LPOS信号在磁带速度变化和恒定期间产生。 [0023] Preferably the present invention provides a computer program product, wherein the LPOS signal changes in tape speed and constant during production.

[0024] 本发明优选提供一种计算机程序产品,其中,用于LPOS探测的近似最大速度由Vmax = Lp女(fs/2)来表示,其中,Lp是所述伺服脉冲图形中的磁性跃变之间的最小距离, 单位为微米,而fs为所述固定采样率,单位为MHz。 [0024] Preferably the present invention provides a computer program product, which is used to approximate maximum speed LPOS detection by Vmax = Lp female (fs / 2) to indicate where, Lp is the servo burst pattern of magnetic transitions The minimum distance between units is m, and fs is the fixed sampling rate, in units of MHz.

附图说明 Brief Description

[0025] 图1显示了嵌有LPOS信息的伺服脉冲的LTO规格; [0025] Figure 1 shows the embedded servo bursts LTO specification of LPOS information;

[0026] 图2显示了现有技术中的LPOS异步探测体系结构的方框图; [0026] Figure 2 shows a block diagram of a prior art LPOS asynchronous detection architecture;

[0027] 图3显示了本发明中的LPOS同步体系结构的方框图; [0027] Figure 3 shows a block diagram of the present invention LPOS synchronous architecture;

[0028] 图4显示了可以实现本发明的同步伺服信道的方框图; [0028] FIG. 4 shows a block diagram can be synchronized servo channel of the present invention;

[0029] 图5是描述初始获取过程的流程图; [0029] FIG. 5 is a flowchart describing the initial acquisition process;

[0030] 图6画出了用于计算磁带速度和y位置估计值的峰到达时间之间的间隔; [0030] FIG. 6 shows the tape velocity and y is used to calculate the estimated value of the peak position of the interval between arrival;

[0031] 图7画出了正确获取的概率; [0031] Figure 7 shows the probability of correct acquisition;

[0032] 图8画出了错误获取的概率; [0032] Figure 8 shows the probability of a false acquisition;

[0033] 图9画出了在伺服信道输出中速度误差的归一化标准偏差与信噪比的关系; [0033] FIG. 9 depicts the servo channel output speed error of normalization of the relationship between the standard deviation and SNR;

[0034] 图10画出了在伺服信道输出中平均获取时间与信噪比的关系; [0034] Figure 10 shows the average acquisition time and signal to noise ratio in relation servo channel output;

[0035] 图1IA画出了双位脉冲的自关联; [0035] FIG 1IA depicts the autocorrelation double-pulse;

[0036] 图IlB画出了所述自关联函数的偏微分; [0036] FIG IlB depicts the autocorrelation function of partial differential;

[0037] 图12显示了时序基准产生系统的操作的方框图; [0037] Figure 12 shows a block diagram of the timing basis generation of the operating system;

[0038] 图13显示了时序基准产生系统的更详细的方框图; [0038] Figure 13 shows a more detailed block diagram of the timing basis generation system;

[0039] 图14显示了时序基准产生系统的线性等价模型;[0040] 图15显示了时序基准产生系统的简化模型; [0039] Figure 14 shows the linear equivalent model of the timing basis generation system; [0040] Figure 15 shows a simplified model of the timing basis generation system;

[0041] 图16A和16B分别显示了对于ν = 0. 5m/s和ν = 12. 5m/s的插值信号样本和时序相位收敛性; [0041] FIG. 16A and 16B show that for ν = 0. 5m / s and ν = 12. 5m / s of the interpolated signal sample and timing phase convergence;

[0042] 图17是描述异步监视功能块的操作的流程图; [0042] FIG. 17 is a flowchart describing the operation of the asynchronous monitoring function block;

[0043] 图18是一个时序图,显示了观察窗口的产生; [0043] FIG. 18 is a timing diagram showing the generation of observation window;

[0044] 图19显示了同步监视功能块的方框图; [0044] Figure 19 shows a block diagram of the synchronous monitoring function block;

[0045] 图20A和20B分别显示了对于ν = 6m/s和ν = 12m/s的y位置估计值的平均值和标准偏差; [0045] FIG. 20A and 20B show the position estimate for y ν = 6m / s and ν = 12m / s of the mean and standard deviation;

[0046] 图21A和21B分别显示了对于ν = 6m/s和ν = 12m/s的磁带速度估计值的标准 [0046] Figures 21A and 21B show that for ν = 6m / s and ν = 12m / s in the standard tape velocity estimate

偏差; Deviation;

[0047] 图22分别是基于峰探测的LPOS符号的大数判决解码规则(majoritydecoding rule)的示意图和表格; [0047] FIG. 22 are based on the detection of the peak of Tarsus LPOS symbol decision decoding rule (majoritydecoding rule) schematic diagram and tables;

[0048] 图23显示了最佳LPOS探测系统的方框图; [0048] Figure 23 shows a block diagram of the optimum LPOS detection system;

[0049] 图24A和24B分别显示了用于A、B脉冲中的第二和第四双位中的度量计算的匹配滤波器波形; [0049] FIG. 24A and 24B show the matched filter waveform for A, B pulses a second and fourth place in the double-metric calculation;

[0050] 图25显示了简化的LPOS符号探测系统的方框图; [0050] Figure 25 shows a block diagram of a simplified LPOS symbol detection system;

[0051] 图26A和26B分别画出了在磁带加速期间所估计出的速度和所计算出的度量值; 以及 [0051] FIG. 26A and 26B are shown in the tape during acceleration the estimated speed and the calculated measure; and

[0052] 图27显示了用于产生SDR和符号判决可靠性估计值的系统的方框图。 [0052] FIG. 27 shows a block diagram of a system for generating a symbol decision SDR and reliability of estimates. 具体实施方式 DETAILED DESCRIPTION

[0053] 同步伺服信道体系结构 [0053] The synchronous servo channel architecture

[0054] 在本发明所述的体系结构中,动态插值器允许在探测器输入中采用任何采样频率,唯一的限制来源于潜在的混叠效应(aliasing effect) 0例如,在磁跃变之间的距离为Lp = 2. 1 μ m的情形中,在15MHz固定ADC时钟频率下不发生混叠效应的可达到的最大磁带速度为Vmax = 2. IX (15/2) = 15. 75m/s。 [0054] In the architecture of the present invention, the dynamic interpolation allows the use of any input sampling frequency detector, the only restriction comes from the potential effects of aliasing (aliasing effect) 0, for example, in magnetic transitions between distance of Lp = 2. 1 μ m of cases, does not occur in a fixed 15MHz clock frequency ADC aliasing effects can reach a maximum speed of tape Vmax = 2. IX (15/2) = 15. 75m / s . 在这种情形中,在脉冲插值器之后用于实现0.05 μ m分辨率的等价采样频率等于315MHz。 In this case, after the pulse interpolator sampling frequency is used to achieve an equivalent resolution of 0.05 μ m is equal 315MHz. 注意,这将是异步体系结构中所要求的固定采样频率。 Note that this will be fixed sampling frequency in an asynchronous architecture required.

[0055] 图3显示了本发明所述的同步LPOS探测结构400的基本构造框图。 [0055] Figure 3 shows a block diagram of the basic configuration of the present invention, the synchronous LPOS detection architecture 400. 因为在脉冲插值器之后单位长度的样本数是固定的且不依赖于磁带速度,所以可以使用匹配滤波器方法以便在有噪声的情况下最佳地探测PPM。 Because the number of samples after pulse interpolator unit length is fixed and does not depend on the speed of the tape, it is possible to use a matched filter method in the presence of noise detected best PPM. 所述匹配滤波器具有固定的单位长度样本数。 The matched filter has a fixed number of samples per unit length. 此外,通过测量基于时序的伺服脉冲的单个双位的过零点之间的距离,可以获得位置误差信号和所述速度估计值。 Further, by measuring over a single dibit pulse timing based servo zero distance between the position error signal can be obtained and the speed estimated value. 在斜升斜降(ramp-up,ramp-down)期间,对于任何小于等于最大磁带速度的恒定磁带速度,能够产生可靠的位置误差信号、获得可靠的磁带速度估计和进行可靠的LPOS探测,其中最大磁带速度的初步近似为 In the ramp-down ramp (ramp-up, ramp-down) period, is less than equal to the maximum for any constant tape speed tape speed, can produce reliable position error signal, to obtain reliable tape velocity estimates and reliable LPOS detection, wherein The maximum tape speed as a first approximation

[0056] Vfflax = LpX (fs/2) (1) [0056] Vfflax = LpX (fs / 2) (1)

[0057] 其中,Lp为磁性跃变之间的最小距离,单位为μ m,而fs为ADC的固定采样频率,单位为MHz。 [0057] where, Lp is a minimum distance between magnetic transitions in units of μ m, and fs is fixed ADC sampling frequency in MHz. 对用于伺服信道选择的信号与噪声加失真之比进行的监测也能够得到支持。 Monitoring than for servo channel selection signal and noise plus distortion of conduct can also be supported.

[0058] 图4显示了同步伺服信道400的更详细的方框图。 [0058] FIG. 4 shows a more detailed block diagram of synchronous servo channel 400. 由获取电路402来获取同步伺服信道参数的初始值。 By the acquisition circuit 402 to obtain the initial value of the synchronous servo channel parameters. 由参数估计块404提供磁带速度和y位置估计以及监测伺服信道操作。 Parameter estimation block 404 provides an estimated tape velocity and y position as well as monitoring the servo channel operation. 由时序基准产生块1300确定必须产生插入信号样本以便以预定的固定频率独立于磁带速度来获得这些样本的时刻。 Generated by the timing reference block 1300 to determine the signal samples must be inserted so as to generate a predetermined fixed frequency independent of the speed of the magnetic tape to obtain the time of these samples. 最后,由最佳LPOS探测块2300来产生LPOS符号判决。 Finally, the best LPOS detection block 2300 to generate LPOS symbol decisions.

[0059] 同步伺服信道参数的初始获取 [0059] Initial synchronous servo channel parameters acquisition

[0060] 设计同步伺服信道的一个主要挑战是,为作为信号插值基础的时序基准的产生确定初始时刻和标称步插值间隔(nominal step interpolationinterval)。 [0060] A major challenge in designing synchronous servo channel that is used as the basis of a timing reference signal interpolation generated to determine the initial time and the nominal step interpolation interval (nominal step interpolationinterval). 标称步插值间隔Ti (单位为微秒)为磁带走过标称步插值距离Xi (单位为微米)所用的时间,它由下式给出 The nominal step interpolation interval Ti (microseconds) for the tape through the nominal step interpolation distance Xi (in microns) The time, which is given by the following formula

[0061] Ti (2) [0061] Ti (2)

ν ν

[0062] 其中,ν表示磁带速度(单位为m/s)。 [0062] where, ν represents the tape speed (m / s). 此外,对于基于匹配滤波的最佳LPOS符号探测,也需要知道伺服读磁头的横向(y)位置信息。 In addition, for optimum LPOS symbol detection based on matched filtering, also we need to know the servo read head transverse (y) location. 所以,作为初始获取过程的结果,除了初始时刻和标称步插值间隔之外,还需要获得可靠的磁带速度和y位置的估计。 So, as a result of the initial acquisition process, in addition to the initial time and the nominal step interpolation interval, also you need to get a reliable estimate of the tape velocity and y position.

[0063] 应该使用模拟_数字转换器(ADC)输出信号样本序列在事先没有任何磁带速度和y位置信息的情况下进行初始获取。 [0063] _ should use the analog to digital converter (ADC) output signal sample sequence in advance without any tape velocity and y position information to conduct initial acquisition. 所以,重要的是设计一种方法,可以在很宽的磁带速度范围内(在LTO磁带驱动系统中一般在0.5m/s到12.5m/s的范围内)完成信道参数的可靠的初始获取。 Therefore, it is important to design a method, a wide range of tape speeds (in LTO tape drive system typically 0.5m / s in the range of 12.5m / s to) completion of channel parameters reliable initial acquisition. 由于ADC的采样频率是固定的,通常在15MHz到24MHz的范围内,所以当磁带速度低时,对于伺服脉冲的每个双位可以获得若干样本,而当磁带速度高时,对于伺服脉冲的每个双位只能提供给获取电路很少几个样本。 Since the ADC sampling frequency is fixed, usually in the range of 15MHz to 24MHz, so when the tape speed is low, can be obtained for a number of samples per dibit servo bursts, and when the tape speed is high, the servo pulses per a double-capture circuit can only provide very few samples. 由于在初始获取开始时没有时序信息, 所以,所述获取方法依赖于对伺服脉冲的双位峰的观测。 Since the initial acquisition start no timing information, so the acquisition method relies on observations of the servo burst dibit peaks. 通常,如果信道输出信号样本的绝对值超过了给定的阈值,那么,可以探测到伺服脉冲中的双位的正的或负的峰。 Typically, if the absolute value of the channel output signal sample exceeds a given threshold, then, it can detect the servo burst of positive or negative peak double bits. 所以,在低速的情形中,所述获取方法必须能够探测到唯一的峰,即使来自单个双位的几个相继的信号样本都超过了阈值。 Therefore, in the case of the low speed, the acquisition method must be able to detect a unique peak even though several bits from a single pair of successive signal samples exceed the threshold value. 在高速的情形中,所述获取方法必须能够处理这样的情形,即来自双位的正的或负的峰的信号样本都没有超过阈值。 In the case of a high speed, the acquisition method must be able to handle such a situation, that is, from the positive or negative peak signal samples are not dibit exceed the threshold. 注意,伺服帧通过C、D、A和B伺服脉冲序列来确定,如图1所示。 Note, the servo frame by C, D, A, and B servo burst sequence is determined as shown in FIG.

[0064] 本发明也基于对伺服帧中有效的[4 4 5 5]脉冲序列的识别来进行同步伺服信道参数的初始获取。 [0064] The present invention is also based on an effective servo frame [4455] Recognition of the pulse sequence for initial synchronous servo channel parameters accessible. 这一点可以通过对伺服脉冲中的双位的正负峰之间的时间间隔进行粗略估计来实现。 This can be time-servo burst through the double-digit gap between positive and negative peak achieved a rough estimate.

[0065] 图5显示了初始获取方法的流程图。 [0065] Figure 5 shows a flow chart of the initial acquisition method. 在获取过程开始的时候,对下列变量和数组进行初始化(步骤500): In the acquisition process began, the following variables and arrays are initialized (step 500):

[0066] a)acqFlag = 0,获取标记指示同步伺服信道处于获取模式; [0066] a) acqFlag = 0, obtain numerals synchronous servo channel is in acquisition mode;

[0067] b) k = 0,在每个采样时刻计数器增加1 ; [0067] b) k = 0, counter is incremented at each sampling time 1;

[0068] c)n = 0,每次信号样本的绝对值超过给定阈值(由thres表示)时,指数增加1 ; [0068] c) n = 0, the absolute value of each signal sample exceeds a given threshold value (denoted by thres), the index is incremented by one;

[0069] d)Np = 0,在一个伺服脉冲中探测到的双位峰(包括正的和负的)的数目; [0069] d) Np = 0, detected in a double-digit peak servo burst (including the number of positive and negative); and

[0070] e) Tp = Tp,max,对一个双位的两个峰之间的时间间隔的粗略估计;初始值Tp,max 的选择使得对于给定范围内的所有磁带速度双位峰之间的时间间隔小于Tp,max ; [0070] e) Tp = Tp, max, a rough estimate of the time a two-bit interval between two peaks; the initial value Tp, max is chosen such that for a given range of the tape speed all the time between the peaks dibit interval is less than Tp, max;

[0071] f). = □,矢量,其元素由峰的到达时间给出; [0071] f) = □, vector whose elements are given by the peak arrival time;

[0072] g)p = □,矢量,其元素由伺服脉冲内所探测到的峰的数目给出。 [0072] g) p = □, vector, whose elements are given by the number of servo pulses detected peaks.

[0073] 在每个采样时刻,计数器增加1(步骤502)并与给定的maxCnt值相比较(步骤504):如果该计数器超过了maxCnt,那么暂停时间(timeoutperiod)就结束(步骤506),重新开始获取过程。 [0073] In each sampling time, the counter is incremented by one (step 502) and with a given maxCnt compared to the value (step 504): if the counter exceeds the maxCnt, then pause time (timeoutperiod) ended (step 506). re-start the acquisition process. 否则,时刻k时的信号样本的绝对值(用|rk|表示)就与给定的阈值进行比较(步骤508)。 Otherwise, the absolute value of the signal sample at time k (using | rk | shown) on a given threshold value (step 508). 如果|rk|超过了所述阈值,那么很可能已经获得了与双位峰(正的或负的)对应的信号样本。 If | rk | exceeds the threshold value, then it may have been obtained with the double-peak (positive or negative) signal corresponding samples. 在这种情形中,指数η增加1,并且将第η个峰的到达时间和信号样本分别存储为& = 1^和4 = 1~1;(步骤510),其中,T表示固定的采样间隔。 In this case, the index η increased by 1, and the first arrival time and the signal samples are stored η peaks is & = 1 ^ = 1 and 4 to 1; (step 510), where, T represents a fixed sampling interval . 如上所述, 需要判断绝对值超过了阈值的信号样本是已经探测到的峰还是新的峰。 As mentioned above, we need to determine the absolute value exceeds the threshold value of the signal sample is already detected peak or a new peak. 为此,将rn的符号与绝对值超过了阈值的前一个样本的符号(由SgnOv1)表示)进行比较(步骤512)。 For this purpose, the sign of the absolute value of rn exceeds the previous symbol sample represents the threshold (manufactured by SgnOv1)) (step 512).

[0074] 如果sgn(rn) Φ SgnOv1),那么这两个样本可能属于不同的峰。 [0074] If sgn (rn) Φ SgnOv1), then the two samples may belong to different peaks. 将这两个样本之间的时间间隔k-tM与双位峰之间的时间间隔的当前估计相比较(步骤514)。 These two current estimate of the time interval between samples and the k-tM interval between dibit peaks is compared (step 514). 如果Tp > k-ty,那么,对估计值进行更新,赋予新值Tp = H1 (步骤516)。 If Tp> k-ty, so, to estimate the value to be updated, giving new value Tp = H1 (step 516). 注意,在获取过程期间, Tp朝着双位峰之间的时间间隔的期望估计值单调下降。 Note that during the acquisition process, Tp double-digit peak towards the desired time interval between the estimated value decreases monotonically. 由于tn为新峰的到达时间,所以, 其值被附加到矢量τ上,并且对峰的计数Np增加1(步骤518)。 Since tn is the arrival time of a new peak, therefore, its value is appended to the vector τ, and to increase a peak count Np (step 518). 此时,需要判断所述新峰是属于当前正在考虑中的伺服脉冲的一个双位还是属于新脉冲的一个双位。 At this point, we need to determine whether the new peak is part of the current being considered a double-servo burst or pulse of one pair belong to the new position. 为此,将时间差Hlri与值mTp相比较(步骤520),其中,m为常数,设Tp等于一个双位的两个峰之间的标称间隔,m的选择使得对于所有的磁带速度和伺服读磁头的所有的横向位置,在同一脉冲中的相继双位的峰之间的时间间隔小于mTp,而在不同脉冲中的相继双位的峰之间的时间间隔大于mTp。 To this end, the time difference value mTp Hlri and compares (step 520), wherein, m is a constant, Tp set equal to the nominal dibit interval between two peaks, m is chosen such that for all tape speed and servo read All the lateral position of the head, in the same pulse peak time between successive intervals of less than double-digit mTp, but in a different pulse peak time between successive two-bit interval is greater than mTp. 如果tn-tn_i > mTp,那么就探测到一个新的脉冲,将上一个脉冲中所探测到的峰的数目Np-I附加到矢量ρ上,并将当前脉冲的峰计数Np初始化为1 (步骤522)。 If tn-tn_i> mTp, then it detects a new pulse, the last pulse in the detected peak number Np-I appended to the vector ρ, and the peak current pulse count Np initialized to 1 (step 522). 接下来就要验证矢量P的最后四个元素是否对应着一个伺服帧的伺服脉冲中的峰的个数的预期序列,该序列为[8 8 1010](步骤524)。 Next step is to verify whether the last four elements of vector P corresponding to the servo bursts of a servo frame sequence number expected peaks, the sequence [881010] (step 524). 如果验证结果是肯定的,那么,就可以识别与伺服脉冲序列[CDAB]相对应的脉冲序列[4 4 5 5](参见图1),并可以计算启动同步伺服信道操作的初始化参数(步骤526)。 If the verification result is positive, then, we can identify with the servo pulse sequence [CDAB] corresponding pulse sequence [4455] (see Figure 1), and calculate start synchronous servo channel operation initialization parameters (step 526 ). 否则,获取过程继续进行,处理下一个采样间隔(步骤502)。 Otherwise, the acquisition process continues, with the next sampling interval (step 502).

[0075] 如果sgn (rn) = sgn Ov1),那么必须考虑两种情形。 [0075] If sgn (rn) = sgn Ov1), you must consider two scenarios. 如果Tp彡W1 (步骤528), 那么很可能这两个样本属于相同的峰,因此不更新矢量τ和ρ以及变量Νρ,并且获取过程继续进行,处理下一个采样间隔(步骤502)。 If Tp San W1 (step 528), then it is possible that the two samples belong to the same peak, and therefore does not update vectors τ and ρ and variable Νρ, and the acquisition process continues, with the next sampling interval (step 502). 如果Tp < tn-tn_i,那么很可能这两个样本是从属于相继双位的具有同一极性的峰获得的,发生了漏测峰的事件。 If Tp <tn-tn_i, it is likely that the two samples are subordinate consecutive double-digit peaks obtained with the same polarity, leakage test peak events. 在这种情形中,将两个峰到达时间值附加到矢量τ上,用tn'和&表示,并将峰计数Np增加2 (步骤530)。 In this case, the two peak arrival time values appended to the vector τ, with tn 'and & representation and increase in the peak count Np 2 (step 530). 漏掉的峰的到达时间tn'的估计值根据漏掉峰的极性、双位峰之间的时间间隔的估计值Tp、和磁带的移动方向等信息来获得。 Estimate the arrival time of the peak missing tn 'value Tp, and the moving direction of the tape and other information based on the estimated time missed peak polarity dibit interval between the peaks is obtained. 如果磁带向前移动,那么在伺服信道的输出端,首先获得的是具有正极性的双位峰,而如果磁带反向移动,那么首先获得的是具有负极性的双位峰。 If the tape is moving forward, then the output of the servo channel is first obtained having a positive polarity of a double-peak, and if the tape is the reverse movement, then the first double-digit gain is a peak with a negative polarity. 所以, 如果磁带正在向前移动并且漏掉的峰的极性为负,或者如果磁带正在反向移动并且漏掉的峰的极性为正,那么,漏掉的峰的到达时间就可以估计为tn' = U+Ί;,否则,就估计为tn' =tn-Tp。 Therefore, if the tape is moving forward and polar peak missing is negative, or if the tape is moving backward and the polarity of the missed peak is positive, then missed the peak arrival time can be estimated as tn '= U + Ί ;, otherwise, it is estimated at tn' = tn-Tp. 如上所述,此时,需要判断新峰是属于当前正在考虑的伺服脉冲中的一个双位还是属于新脉冲中的一个双位(步骤532)。 As mentioned above, this case, we need to determine a new peak belongs servo burst currently considering a two-position or belonging to a new pulse of a two-position (step 532). 如果k-tM〉!!!!;,那么就探测到一个新脉冲,并且再次需要考虑磁带的移动方向来确定下一步(步骤534)。 If the k-tM> !!!!;, then a new pulse is detected, and the need to consider the direction of movement of the tape again to determine the next step (step 534). 在图5的流程图中,假设磁带向前移动。 In the flowchart of FIG. 5, it is assumed the tape moves forward. 类似的过程可以应用到反向的情形中。 A similar process can be applied to reverse the situation. 如果探测到的峰具有正极性,那么这就很可能对应着所述新脉冲中的第一个峰。 If the probe has a positive polarity peak, then this is likely to correspond to the new pulse of the first peak. 然后,将值Np-I附加到矢量ρ上,并将当前脉冲的峰计数Np初始化为1(步骤536)。 Then, the value of Np-I appended to the vector ρ, and the peak count Np initialization pulse current is 1 (step 536). 然后检查矢量ρ的最后四个元素是否对应着序列[8 8 10 10](步骤538)。 Then check the last four elements of the vector ρ corresponds with the sequence [881 010] (step 538). 另一方面,如果所探测到的峰具有负极性,那么很可能所述新脉冲的第一个峰被漏掉了。 On the other hand, if the detected peaks have negative, then the first of the new pulse peak is likely to be missed. 然后,将值Np_2附加到矢量ρ上,并将当前脉冲的峰计数Np初始化为2(步骤540)。 Then, the value Np_2 appended to the vector ρ, and initialize the peak current pulse count Np is 2 (step 540). 注意,在这种情形中,不检查矢量ρ的最后四个元素是否对应着序列[8 8 10 10],因为在探测到“C脉冲”(参见图1)中的第一个双位的第一个峰时,认为发生了从“获取”模式向“跟踪”模式的转换,从而减小获取系统实现的复杂性。 Note that, in this case, does not check whether the last four element vector ρ corresponds to the sequence [881 010] because the first double-digit in the probe into the "C Pulse" (see Figure 1) in the first When a peak that occurred to the "tracking" mode of transition from "access" mode, thereby reducing acquisition system implementation complexity.

[0076] 一旦在矢量ρ的最后四个元素中探测到序列[8 8 10 10]时,就计算启动同步信道操作所需要的各个参数,并将变量acqFlag设置为1 (步骤526)以表示获取过程成功完成。 [0076] Upon detection of the sequences in the last four elements of the vector ρ [881,010], the purpose of calculating the parameters of the synchronization channel started operating needs, and variable acqFlag set to 1 (step 526) to indicate Gets It was completed successfully. 假设考虑C、D和A脉冲的峰到达时间来计算磁带速度和y位置估计值,如图6所示,则预期的估计值分别由下式给出(步骤526) Suppose consideration C, D and A pulse peak arrival time to compute the tape velocity and y-position estimates, the estimated value shown in Figure 6, respectively, it is expected that given by (step 526)

[0077]少饥0 —〜π ^ 2 tan(—) 30 [0077] less hungry 0 -~π ^ 2 tan (-) 30

AI + A2 + A3 + A4 50 AI + A2 + A3 + A4 50

Λ Λ

51 + 52 + 53 + 54 I 51 + 52 + 53 + 54 I

(3) (3)

^_ (4) ^ _ (4)

[0078] νρκΙ ο =—— [0078] νρκΙ ο = -

BI + B2 + B3 + B4 BI + B2 + B3 + B4

[0079] 其中,磁带向前运动时1 = ΙΟΟμπι,而磁带反向运动时1 =95μπι。 [0079] wherein, when the tape moves forward 1 = ΙΟΟμπι, while the tape backward motion 1 = 95μπι. 一旦完成了获取过程,就开始进行同步伺服信道操作,其中初始时刻为ty = tn,将标称步插值间隔设置为TVtl = XiA^u,用来产生作为信号插值基础的时序基准(步骤526)。 Upon completion of the acquisition process, they begin to carry out synchronous servo channel operation, wherein the initial time for ty = tn, the nominal step interpolation interval to TVtl = XiA ^ u, the basis for generating a timing reference signal interpolation (step 526) . 该获取过程也可以提供由下式给出的参数: The acquisition process can also be provided by the parameters given by:

[0080] [0080]

^pVest,0 ^ PVest, 0

IntpStepO = IntpStepO =

5 5

Ixi Ixi

[0081] 其中,Lzj表示小于等于ζ的最大整数。 [0081] where, Lzj represents the largest integer less than or equal ζ. 这个参数表示按Xi的倍数来衡量的当前样本与对应着C脉冲中的第一个双位的过零点的样本之间的插值步数的估计值,它将被用于由时序恢复环(timing recovery loop)进行的最初时序调整(first timing adjustment)。 This parameter indicates the current sample according to Xi multiples to measure and estimate the number of steps corresponding to the interpolated samples C pulse in the first double digit between zero crossings, it will be used by the timing recovery loop (timing recovery loop) timing adjustment originally carried out (first timing adjustment).

[0082] 在实现获取电路402时,矢量τ和ρ设定为长度分别为36和4的延迟线。 [0082] In the circuit to achieve access to 402, the vector length is set to τ and ρ are delay lines 36 and 4. 另外要注意,为了对估计值进行计算由电路引入的等待时间应该小于伺服脉冲之间的时间间隔。 Also note that, in order to estimate the value calculated by the circuit introduced waiting time should be less than the time interval between the servo bursts. 此外,为了获得可靠的磁带速度和y位置估计值,需要一个电路以高精度来进行正数之间的除法。 Furthermore, in order to obtain reliable tape velocity and y-position estimates, it requires a circuit for high precision division between positive numbers. 通过应用Newton-Raphson算法利用迭代近似计算除数的倒数,然后利用被除数和除数的倒数之积来获得预期的商,可以满足上述要求。 By applying Newton-Raphson iterative algorithm to calculate the approximate reciprocal divisor, dividend and then use the product of the reciprocal of the divisor to get the desired business, to meet the above requirements. 例如,在所述算法进行了三次迭代之后,除法计算的相对误差小于0. 4%,在四次迭代后,相对误差小于1. 6X 10_5。 For example, after three iterations of the algorithm, dividing the calculated relative error is less than 0.4%, after four iterations, the relative error is less than 1. 6X 10_5.

[0083] 通过仿真研究了所述获取方法的性能。 [0083] The simulation study of the performance of the acquisition method. 采用Lorentz模型,磁带记录信道对单个跃变的响应可以用下式表示: Using Lorentz model, tape recording channel can be expressed as a single response transitions:

Θ(ην) = -~—^-y (6) Θ (ην) = - ~ - ^ - y (6)

[0084] π\ + ( 2vt I [0084] π \ + (2vt I

[0085] 其中,参数PW50/2表示在以速度ν移动的磁带上处于平行于伺服带中心线的直线上的两点之间的距离(单位为微米),在这两点处,伺服读磁头分别产生单个跃变的信道响应的最大值和最大值之半。 [0085] where the parameter PW50 / 2 represented on the magnetic tape traveling at a speed of ν in the distance (in microns) between the straight line parallel to the centerline of the two servo band, in which two points, the servo read head generate half the individual transitions and maximum channel response maximum value. 所以,双位信号脉冲由下式给出 Therefore, the double-pulse signal is given by the following formula

[0086] g(t ;ν) = Θ (t ;ν)-Θ (t-T0 ;ν) (7) [0086] g (t; ν) = Θ (t; ν) -Θ (t-T0; ν) (7)

[0087] 其中,T0 = Lp/v微秒,Lp表示磁性跃变之间的距离。 [0087] where, T0 = Lp / v microseconds, Lp represents the distance between magnetic transitions. 回想一下,对于图1所示的 Recall that shown in Figure 1 for

10伺服帧中的双位来说,Lp = 2. 1 μ m0 10 double-servo frame is, Lp = 2. 1 μ m0

[0088] 图7和图8分别显示了对于磁带速度和给定的峰探测阈值的不同值作为伺服信道输出的信噪比(SNR)的函数的正确获取的概率和错误获取的概率。 [0088] Figures 7 and 8 show the probability and the probability for different values given tape speed and peak detection threshold as signal to noise ratio (SNR) servo channel output function to obtain correct errors acquisition. 如果初始y位置估计的误差的绝对值小于2. 5 μ m,那么就认为是正确的获取,否则就认为是错误的获取。 If the initial estimate of the absolute value of y position error is less than 2. 5 μ m, then that is the right to acquire, or that it is wrong to obtain. 每个概率值由获取过程的500次实现来获得,假设信道为用于产生伺服脉冲的PW50/2. Ium = 0.4 的Lorentz型信道,噪声为加性高斯白噪声、ADC的采样频率为15MHz,磁带速度恒定、以及y 位置为零。 Each probability value from the acquisition process of 500 times to achieve to get, assuming that the channel is used to generate a servo pulse PW50 / 2. Ium = Lorentz-type channel 0.4, the noise is additive white Gaussian noise, ADC sampling frequency is 15MHz, the tape speed is constant, and y position zero. 对于不同的磁带速度值,图9和图10分别显示了由正确获取所给出的初始磁带速度估计值误差的归一化标准偏差(其中归一化因子由磁带速度给出)和平均获取时间。 For different tape speed value, Figures 9 and 10 show the initial tape speed is given by the right to obtain an estimate of the error normalized standard deviation (where the normalization factor is given by the tape speed) and the average acquisition time . 在图9和图10中,对于范围在0. 5m/s到12. 5m/s的磁带速度值画出了所述归一化标准偏差和平均获取时间的四条曲线。 9 and 10, for the range of 0. 5m / s to 12. 5m / s tape speed value depicts the normalized standard deviation and the mean acquisition time of four curves.

[0089] 用于信号插值的时序基准的产生 [0089] The timing reference signal interpolation for generating

[0090] 伺服信道的同步操作要求为信号插值产生时序基准,以便独立于磁带速度以预定的固定频率(每微米1/Xi个样本)获得插入的信号样本,其中,Xi表示标称步插值距离。 [0090] The servo channel synchronous operation is required to produce a timing reference signal interpolation so independently of the tape speed to a predetermined fixed frequency (per micrometer 1 / Xi samples) to obtain signal samples inserted, which, Xi represents the step interpolation distance . 例如,如果Xi = 0. 05 μ m,那么频率1/Xi等于每微米20个样本。 For example, if Xi = 0. 05 μ m, then the frequency 1 / Xi is equal to 20 samples per micrometer. 用于产生时序基准的固有参考由伺服脉冲来提供,伺服脉冲周期性地出现在伺服信道输出中,如图1所示。 Inherent timing reference for generating the reference pulse is provided by a servo servo pulses appear periodically in the servo channel output, as shown in FIG. 所以,伺服信道输出中的信号可以被认为是引导信号(Pilot signal),从中可以提取时序信息。 Therefore, the servo channel output signal can be considered as a pilot signal (Pilot signal), from which you can extract the timing information. 然而,从伺服信道信号中提取时序信息不是直截了当的,因为A、B、C、D伺服脉冲不是等间隔的。 However, the extract from the servo channel signal timing information is not straightforward, because A, B, C, D servo bursts are not equally spaced. 此外,伺服脉冲之间的间距依赖于伺服读磁头的y位置,并且伺服帧的重复周期以及一个伺服脉冲中的相继双位之间的时间间隔依赖于磁带速度。 Moreover, the spacing between the servo bursts depends on the y-position servo read head, and the time and repetition period of the servo frame in a servo burst interval between successive dibit depends on the tape velocity. 此外,需要考虑用于对A和B 脉冲中的LPOS信息进行编码的脉冲位置调制的存在。 In addition, we need to consider for A and B pulses in the presence of information encoded LPOS pulse position modulation. 对于磁带向前移动的情形,伺服信道信号由下式给出: For the case of the tape moves forward, the servo channel signal is given by:

[0091] [0091]

Figure CN101361125BD00111

[0092] [0092]

Figure CN101361125BD00112

[0093] 其中,τ表示要恢复的时序相位,矢量b表示属于二进制字符{0,1}的LPOS符号序列,w(t)为加性高斯白噪声,其谱密度为Ntl,&为伺服信道信号中的伺服帧的数目, Qi ( ), i = 0,...,3 分别表示C、D、A、B 伺服脉冲。 [0093] where, τ represents a timing phase to be restored, vector b represents LPOS symbol sequence belongs to {0,1} of binary characters, w (t) is the additive white Gaussian noise, the spectral density of Ntl, & servo channel The number of frames of servo signal, Qi (), i = 0, ..., 3 respectively C, D, A, B servo bursts. 定义ak = 2bk_l,ak e {-1,+1},则伺服脉冲可以表示为 Definition ak = 2bk_l, ak e {-1, + 1}, the servo bursts can be expressed as

[0094] [0094]

Figure CN101361125BD00113

[0095] [0095]

Figure CN101361125BD00114

[0096] [0096]

Figure CN101361125BD00115

[0097] [0097]

Figure CN101361125BD00121

[0098] 其中,Tf = Lf/v微秒,Td = Ld/v微秒,ξ = 0. 05,以及g(t ;ν)表示双位信号脉冲,它依赖于磁带速度ν,如公式(7)所定义的。 [0098] where, Tf = Lf / v microsecond, Td = Ld / v microsecond, ξ = 0. 05, and g (t; ν) represents dibit signal pulse, which depends on the tape velocity ν, as shown in equation ( 7) as defined above. 回想一下,Lf = 200ym,Ld = 5μπι,如图1 所示。 Recall, Lf = 200ym, Ld = 5μπι, shown in Figure 1. 对于磁带反向移动的情形,在A和B脉冲的定义中乘以Tf项的系数分别选为95/200 和145/200以替代1/2和3/4的条件下,信号r(t)可以以类似于⑶的方式来表示。 For the case of reverse movement of the tape, is multiplied by the item definition Tf A and B pulses are coefficients selected to 95/200 and 145/200 instead of under the conditions of 1/2 and 3/4, the signal r (t) ⑶ may be similar way to represent.

[0099] 为了估计时序相位,经典的时序恢复方法首先确定似然函数,然后就不想要的随机变量对其进行平均,最后求出使所得到的函数最大化的时序相位的值。 [0099] In order to estimate the timing phase, classical timing recovery method first determines the likelihood function, then the random variable is not want its average of the last determined value so that the resulting maximum phase timing function. 观察(8)和(9) 可以发现,所述似然函数依赖于时序相位、伺服读磁头的y位置、磁带速度、和LPOS 二进制符号序列。 Watch (8) and (9) can be found, the likelihood function depends on the timing of the phase servo read head y position, tape speed, and LPOS sequence of binary symbols. 于是,似然函数可以表示为: Thus, the likelihood function can be expressed as:

Figure CN101361125BD00122

[0104] 引入脉冲响应为gM(t ;ν) = g(-t ;ν)的匹配滤波器并定义卷积积分h(t ;ν) = r 女g„(t ;ν),则似然函数的表达式就变为: [0104] introduction impulse response gM (t; ν) = g (-t; ν) and the definition of the matched filter convolution integral h (t; ν) = r F g "(t; ν), the likelihood function of the expression becomes:

Figure CN101361125BD00123

[0108] 假设y位置和磁带速度的联合概率分布以及LPOS符号的先验分布已知,那么,时序相位的最大似然(maximum-likelihood,ML)估计就由下式给出: [0108] hypothesis y-position and tape velocity joint probability distribution and LPOS symbol prior distribution is known, then the timing of the phase of maximum likelihood (maximum-likelihood, ML) is estimated to is given by the formula:

[0109] [0109]

Figure CN101361125BD00131

[0110] 然而,应用经典的最大似然估计方法来获得时序相位的估计会有如下的严重困难: [0110] However, the serious difficulties the classical maximum likelihood estimation method to estimate the timing of the phase will be obtained as follows:

[0111] a)上述公式中的最大似然估计方法不能很好地用于直接实现,甚至不能很好地用于以时序恢复反馈环的形式(每个伺服帧都计算误差项)来实现,因为计算的复杂性太大。 [0111] a) the above formula maximum likelihood estimation method does not work well for direct implementation, can not even work well for timing recovery in the form of a feedback loop (each calculated servo frame error term) to achieve, Because the computational complexity too.

[0112] b)求似然函数的最大值要求首先所述计算卷积积分相对于τ的偏微分;这个操作可以在数字域中进行,但通常会导致较大的实现复杂性和不可忽略的噪声增强。 [0112] b) seeking the maximum likelihood function first requires the calculation of the convolution integral partial differential with respect to τ; this operation can be performed in the digital domain, but generally result in large implementation complexity and can not be ignored noise enhancement.

[0113] c)似然函数所依赖的随机变量可以具有时变概率分布;例如,在斜升斜降期间所出现的磁带速度的变化会导致在几个伺服帧内磁带速度和时序相位的联合概率分布有显著的变化。 [0113] c) the likelihood function depends random variables may have varying probability distribution; for example, vary during ramp up and ramp down the tape speed appears several servo frame will result in tape speed and timing phase of joint a significant change in the probability distribution.

[0114] 本发明还基于时序恢复环产生用于同步伺服信道操作的时序基准,在所述时序恢复环中,通过观察插入的伺服信道信号的过零点来确定时序调整。 [0114] The present invention is also based on the timing recovery loop generates a timing reference for the synchronous servo channel operation, the timing recovery loop, by observing the insertion of the servo channel signal zero crossings to determine the timing adjustment. 假设可以有效地计算y 位置和磁带速度的可靠估计j)和“如在下一部分中所示的,并且注意到在似然函数(11)中指数上的求和里只有四项依赖于LPOS符号ak (对于每个k),于是时序相位的估计可以近似为: Assumptions can efficiently compute y-position and tape velocity reliable estimate j) and "as shown in the next section, and noted that the summation index on the likelihood function (11), only four rely on LPOS symbol ak (for each k), then the timing phase estimation can be approximated as:

[0115] “ aiS^ Lw{e,y,v), ( ) [0115] "aiS ^ Lw {e, y, v), ()

[0116]其中, [0116] where

[0117] [0117]

Figure CN101361125BD00132

[0119] 在上述公式中,去掉了似然函数对LPOS符号序列的依赖,以及就y位置和磁带速度进行的平均。 [0119] In the above formula, remove the likelihood function dependence on LPOS symbol sequences, as well as on y-position and tape speed average. 然而,仍然需要计算相对于时序相位的偏微分以便求出似然函数的最大值。 However, still need to calculate partial derivatives with respect to the timing phase to find the maximum likelihood function. 引入记号:P和ΐ来表示y位置和磁带速度的估计值以替代在(3)和(4)中定义的yest和Vest, 因为采用了低通滤波来产生j^Pf,而yest和是指即时估计。 Introducing notation: P and ΐ to represent the estimated value y-position and tape velocity to replace in (3) (4) and the definition of yest and Vest, because the use of the low-pass filter to generate j ^ Pf, and yest and means immediate estimate.

[0120] 注意,对于中等的到大的信噪比的值,由双位脉冲的自关联函数Cg( τ -Ts ;ν)= g ^ gM(x-Ts;v)的峰来近似确定h( τ ;ν)的每个峰(恰当选择时移Ts)。 [0120] Note that the value for medium to large signal to noise ratio of the autocorrelation function Cg double-pulse (τ -Ts; ν) = g ^ gM (x-Ts; v) determine the peak to approximate h (τ; ν) of each peak (shift Ts appropriate selection). 另外注意,由于g(T ;ν)是一个具有奇对称的函数,所以自关联Cg(x ;ν)为具有偶对称的函数。 Also note that, due to the g (T; ν) is an odd function with symmetrical, so autocorrelation Cg (x; ν) as a function of having even symmetry. 所以,对于时序相位相对于最佳值的小偏差,h(x ;ν)对于时序相位的偏微分,从而似然函数Z^v(Gj)J)的偏微分,就具有奇对称。 Therefore, the timing of phase with respect to small deviations optimum value, h (x; ν) for partial differential phase sequence, so the likelihood function Z ^ v (Gj) J) of the partial derivatives, it has odd symmetry. 这意味着,在所述双位脉冲被认为是时序相位相对于最佳值的偏差的函数的条件下,如图IlA和IlB所示,h( τ ;ν)的偏微分的行为类似于双位脉冲在自变量很小时的行为。 This means that, in the double-pulse sequence is considered to be a function of the deviation of the phase with respect to the optimum value of the condition, as shown in FIG IlA and IlB, h (τ; ν) is similar to the behavior of the partial derivative bis the argument very hour pulse behavior.

13[0121] 还可以注意到,所获得的h( τ ;ν)的峰对应着伺服信道信号的过零点,该过零点出现在伺服脉冲中的双位峰之间,于是可以得出结论,最大似然方法所给出的时序信息近似等于通过直接在对应着过零点处对伺服信道信号进行采样所获得的时序信息。 13 [0121] may also be noted that the obtained h (τ; ν) peak corresponds to the servo channel signal zero crossing, which had appeared in the servo burst between zero double-bit peak, so can be concluded that the maximum likelihood timing information given is approximately equal to direct the corresponding zero-crossing point of the servo channel signal sampling timing information obtained. 所以,假设f表示时序相位的可靠估计,即f = 那么,预期的时序信息就可以表示为: So, assuming that f represents a reliable estimate of the timing phase, namely f = So, the expected timing information can be represented as follows:

Figure CN101361125BD00141

其中,Γ表示增益因子,巧表示加性高斯白噪声(additive whiteGaussian noise, AffGN)样本。 Where, Γ represents a gain factor, coincidentally represents additive white Gaussian noise (additive whiteGaussian noise, AffGN) samples. 注意,完全避免了由偏微分的计算所导致的噪声增强。 Note, completely avoid the noise enhancement by the resulting partial differential calculation. 在这种情形中, 在模拟_数字转换之前所进行的抗混叠滤波操作近似等价于用来为信号Ht)产生足够的统计所进行的理想滤波。 In this case, the anti-simulation _ before the digital conversion performed by aliasing filtering operation is approximately equivalent to a signal Ht) sufficient to produce the desired statistical filtering performed.

[0127] 这样就可以设计一种时序基准产生系统1200,如图12所示,该系统依赖于被插入的伺服信道信号的过零点来确定误差反馈配置中的时序误差。 [0127] This allows the design of a timing basis generation system 1200, shown in Figure 12, the system relies on the inserted servo channel signal zero crossings to determine the timing error feedback configuration errors. 注意,与时序误差表达式(15)中的求和相对应的平均操作由环滤波器(loop filter) 1202来进行,假设该滤波器具有比例积分器(proportional-plus-integrator)类型。 Note that the timing error in the expression (15) corresponds to the sum of the average operations by the loop filter (loop filter) for 1202, assuming that the filter has a proportional-integral control (proportional-plus-integrator) type. 所以,设有可靠的y位置和磁带速 So, with a reliable y-position and tape velocity

度估计以及可忽略的插值误差,那么,输入到环滤波器的时序误差估计<、)由下式给出 Estimation and interpolation error can be ignored, then the input to the loop filter timing error estimate <,) is given by the following formula

[0128] [0128]

Figure CN101361125BD00142

[0129] 其中,、表示由时序基准产生系统提供的在用于时序恢复的伺服信道信号的过零点附近进行信号插值的时刻,\表示、对过零点时刻的偏差,\为AWGN样本。 [0129] ,, which represents a time reference is generated by the timing system in the vicinity of a timing recovery of the servo channel signal zero-crossing signal interpolation, \ said zero crossing time of deviation, \ is AWGN samples.

[0130] 时序基准产生系统1300的方框图示于图13。 [0130] The block diagram of the timing basis generation system 1300 in FIG. 13. 序列{tn}表示确定信号样本{r (tn)} 的时刻,使得这些样本以预期的固定频率(每微米1/Xi个样本)来获得,而不依赖于磁带速度。 Determining the sequence {tn} indicates signal samples {r (tn)} time, so that these samples to the desired fixed frequency (per micrometer 1 / Xi samples) to obtain, without depending on the tape speed. 很清楚,标称步插值间隔TjPADC采样间隔T 一般是不相称的。 Clearly, the nominal step interpolation interval TjPADC sampling interval T is generally not proportionate. 所以采用线性插值来在插值时刻{tn}获得信号样本,而这些插值时刻通过递归方法来获得: So using linear interpolation in the interpolation time signal samples obtained {tn}, and these moments by recursive interpolation method to obtain:

[0131] tn+1 = tn+T,,n (17) [0131] tn + 1 = tn + T ,, n (17)

[0132] 其中,Ti, n表示由(2)所定义的标称步插值间隔的估计值。 [0132] wherein, Ti, n represents an estimate of the nominal value (2) defined by the step interpolation interval. 标称步插值间隔的估计值由下式给出 The nominal step interpolation interval estimate is given by the following formula

[0133] Tin =Ti^+Atni (18) [0133] Tin = Ti ^ + Atni (18)

[0134] 其中,二为直接从磁带速度估计值所推出的标称步插值间隔的估计值, [0134] One, two direct step interpolation interval estimate of the tape velocity estimates from the launch of the nominal,

[0122] [0122]

[0123] [0123]

[0124] [0124]

[0125] [0125]

Figure CN101361125BD00143

[0126]为在环滤波器输出所获得的校正项,nz为小于等于η的最大时间指数(time index),在该时间指数处时序误差估计值已经被输入到环滤波器中。 [0126] The correction term in the loop filter output obtained, nz is less than the maximum time index equal to η (time index), the time at the timing error estimate the index has been entered into the ring in the filter. 在插值_时间计算单元中,第η个插值时刻被表示为: _ Time interpolation computation unit, the first η a interpolation time is expressed as:

[0135] [0135]

Figure CN101361125BD00151

[0136] 其中,1^和μ η分别表示第η个插值时刻按照采样间隔T的倍数来衡量所得到的整数部分和小数部分。 [0136] where 1 ^ and μ η η represent the first time an interpolation of the sampling interval T is measured in multiples of the integer and fractional parts get. 因此,线性插值就产生了由下式给出的插值信号样本: Accordingly, linear interpolation is produced by the interpolated signal samples given by the following equation:

[0137] Htn^rkJjuh) (20) [0137] Htn ^ rkJjuh) (20)

[0138] 时序基准产生系统1300的目标是提供能够再现由伺服读磁头所产生的信号的插值信号样本,这些插值信号样本对应着位于平行于伺服带中心线的直线上并按标称步插值距离Xi等距隔开的磁带上的点。 [0138] Timing generating target reference system 1300 is to provide interpolated signal samples can be reproduced by the servo read head signal produced, which interpolated signal samples corresponding to the servo band is located parallel to the center line of a straight line on the nominal step interpolation distance press Xi equally spaced points on the tape. 然而,y位置和磁带速度的值通常使得插值信号样本不是精确地对应着伺服信道信号的零点。 However, y values of the position and speed of the tape so that interpolated signal samples are usually not exactly correspond to zeros of the servo channel signal. 换言之,时序误差估计由插值信号样本给出,而该插值信号样本在位于标称步插值距离Xi的倍数处并最接近过零点的点处进行计算,如时序基准产生系统所估计的那样。 In other words, the timing error estimate given by the interpolated signal samples, and the interpolated signal samples located in the step interpolation distance Xi of multiples and over at the point closest to zero are calculated, such as the timing basis generation system as estimated. 所以,标称步插值距离Xi确定了时序误差估计的标准偏差的下限。 Therefore, the step interpolation distance Xi determine the lower limit of the timing error estimate of the standard deviation. 假设过零点位置是一个在标称步插值间隔内均勻分布的随机变量,则时序误差估计的标准偏差的下限就等于。 Assume the position of a zero-crossing point in the lower nominal step interpolation uniformly distributed random variable interval, the timing error estimate is equal to the standard deviation. 例如,如果Xi = 0. 05 μ m,那么过零点位置误差估计的标准偏差的下限就为14. 4nm。 The lower limit, for example, if Xi = 0. 05 μ m, then the zero-crossing position error estimated standard deviation on the 14. 4nm. 注意,原理上,通过减小标称步插值距离Xi的大小能够任意减小这个下限。 Note that, in principle, by decreasing the nominal step interpolation distance Xi of any size can reduce this limit. 实际上,与AWGN对过零点位置误差估计的标准偏差的贡献相比,χ, /(Vl2v) 项小一些就足够了。 In fact, the contribution of the zero-crossing point with AWGN position error estimated standard deviation compared to, χ, / (Vl2v) smaller items is sufficient.

[0139] 如前面所提到的,伺服信道信号的过零点不是均勻隔开的。 [0139] As mentioned previously, the servo channel signal is not uniform over the spaced zero. 具体说,对时序信息产生贡献的过零点之间的时间间隔依赖于伺服读磁头的y位置和磁带速度。 Specifically, the time between the zero crossings of the timing information generated at intervals dependent on the contribution of the servo read head y-position and tape speed. 另外回想一下, 用于LPOS位(LPOS bits)编码的A和B脉冲中的双位的过零点不用于提取时序信息。 Further Recall for LPOS bit (LPOS bits) coded pulses A and B in the two-bit zero crossings are not used to extract timing information. 如图13所示,用于提取时序信息的插值信号样本的选择由称作“控制单元” 1304的有限状态机连同插值“计数器” 1306来完成。 As shown in Figure 13, it is used to select the interpolation signal samples to extract timing information from the called a "control unit" FSM 1304 along with interpolation "Counter" 1306 to complete. 每次计算出一个新的插值信号样本时,“计数器” 1306 就使输入“控制单元” 1304的变量IntpCnt增加1。 Each time a new interpolation calculated signal samples, "counter" 1306 on the input, "control unit" variable IntpCnt 1304 increased by 1. 假设“插值时间计算”单元410所使用的标称步插值间隔Ti的估计值是可靠的,并且关于伺服帧的开始的信息(即一个伺服帧中的C脉冲里的第一个双位的第一个过零点的时刻)是已知的,那么,y位置的可靠估计值的信息对于由“控制单元” 1304来确定必须选择哪一个插值信号样本以提取时序信息是足够的。 If "Interpolation time computation" nominal unit used 410 Ti step interpolation interval estimate is reliable, and the information about the beginning of the servo frame (ie a servo frame C pulse in the first double-digit first a zero-crossing time) is known, then the information is reliable estimate y position of the "control unit" 1304 is determined by the need to choose which interpolated signal samples to extract timing information is sufficient. “控制单元” 1304输出下列变量: "Control unit" 1304 outputs the following variables:

[0140] a) burstFlag,该变量根据0 — C脉冲、1 — D脉冲、2 — A脉冲、3 — B脉冲的规则识别当前伺服脉冲; [0140] a) burstFlag, the variable according to 0 - to identify the current rules servo pulse B pulse C pulse, 1 - D pulse, 2 - - A pulse, 3;

[0141] b)dibitFlag,该变量在伺服脉冲范围内识别当前双位; [0141] b) dibitFlag, this variable to identify the current double-digit range in the servo burst;

[0142] c)neWTimErr,该变量被断言(asserted)以使新的时序误差估计值输入环滤波器; [0142] c) neWTimErr, this variable is asserted (asserted) for the new timing error estimation value of the loop filter;

[0143] cOlntpCntReset,该变量被断言以使插值“计数器” 1306复位。 [0143] cOlntpCntReset, the variable is asserted to enable interpolation "Counter" 1306 is reset.

[0144] 在一个伺服帧的末尾,即在B脉冲中的最后一个双位的过零点处,对变量IntpCntReset进行断言以复位“计数器” 1306,将变量burstFlag置O以表明C脉冲的出现,也将变量dibitFlag置0以表明C脉冲中的第一个双位的出现。 [0144] At the end of a servo frame B pulses that last bit of a double zero crossing, asserts IntpCntReset variables to reset the "counter" 1306, set the variable burstFlag O to signify a C pulse, but also dibitFlag variable is set to 0 to indicate that C pulse in the first two-position appears. C脉冲中的第一个双位与B脉冲中的最后一个双位的过零点之间的插值步数的估计值由下面的变量给出: C pulse in the first two-position and B pulses over the last two-digit number of steps between zero interpolation estimate is given by the following variables:

[0145] [0145]

Figure CN101361125BD00161

[0146] 其中,Ditl表示B脉冲的最后一个双位和C脉冲的第一个双位在y = 0处的距离, 在磁带向前移动和向后移动的情形中它分别等于30 μ m和35 μ m。 [0146] where, Ditl is the last double pulse of B and C of the first double-pulse distance at y = 0, the situation moving forward and backward in the tape it is equal to 30 μ m, respectively, and 35 μ m. 当“计数器” 1306表明IntpCntNum个插值步已经完成时,对变量newTimErr进行断言,相应的插值信号样本作为新的时序误差估计被输入到时序恢复环滤波器1302中,并且将变量dibitFlag置1以表明C脉冲的第二个双位即将出现。 When the "counter" 1306 showed IntpCntNum an interpolation step has been completed, variable newTimErr assertion, the corresponding interpolated signal samples as a new timing error estimate is input to the timing recovery loop filter 1302, and the variable set to indicate dibitFlag The second double-C pulse imminent. 然后,“控制单元” 1304的操作继续进行,计算当前过零点和下一个过零点之间的插值步数。 Then, the operation "control unit" 1304 continues to calculate the current zero-crossing point and the next steps through interpolation between zero. 每次到达一个新的过零点时,对变量newTimErr进行断言并使变量dibitFlag增加1。 Every arrival of a new zero crossing of variables and variable newTimErr assert dibitFlag increased by one. 如果该过零点是某个伺服脉冲的最后一个过零点,那么,使变量burstFlag增加1并将变量dibitFlag置0。 If the zero crossing of a servo burst is the last zero crossing, then the variable is incremented by one and variable dibitFlag burstFlag set to 0. 注意,如果下一个过零点属于同一伺服脉冲,那么,由|_、/x,.」来给出表示过零点之间的估计的插值步数的变量IntpCntNum,如果下一个过零点属于不同的伺服脉冲,那么由类似于(21)的表达式来给出变量IntpCntNum,其中采用当前脉冲的最后一个双位和下一个脉冲的第一个双位在y = 0处的距离来替代Drc, OO如果是用于对LPOS位进行编码的A和B脉冲的第二和第四双位的过零点这种特殊情形, 那么,由IntpCntNum给出的移位了的双位和邻近双位之间的插值步数仍然取为LA/Zx,」。 Note that if the next zero crossing of the same servo burst, then the | _, / x ,. "expressed interpolation step is given the number of variables IntpCntNum estimate of between zero, if the next zero crossing belonging to different servo pulse, then by a similar (21) gives expression to a variable IntpCntNum, where the first two-position using current pulses last double digit and the next pulse of the distance y = 0 to replace Drc, OO if Interpolation is used to encode LPOS bits A and B pulse of the second and fourth double-zero crossing of this particular situation, then shifted by the IntpCntNum given double digit and double digit between neighboring the number of steps still taken as LA / Zx,. " 然而,当计算将与A和B脉冲的第二或第四双位的过零点相对应的插值信号样本时,将dibitFlag增加1而不对newTimErr进行断言。 However, when calculating the A and B pulses of the second or fourth double-digit corresponding to the zero crossings of the interpolated signal samples will increase by 1 instead of newTimErr dibitFlag assert. 一旦达到了B脉冲的最后一个双位的过零点,“计数器” 1306就再次复位,对新的伺服帧的处理就开始了。 Once you reach the B pulses over the last two-bit zero, "counter" 1306 once again reset to the new processing servo frame began.

[0147] 注意,在实际实现用于同步伺服信道的时序基准产生系统1300时,不需要在每个插值步上都清楚地计算插值信号样本。 [0147] Note that in the actual implementation of synchronous servo channel for the timing basis generation system 1300, clearly does not need to calculate the interpolated signal samples for each interpolation step. 如时序调整的产生所要求的,以及脉冲位置调制的LPOS信号(可以在A和B脉冲中遇到)的匹配滤波所要求的,插值信号样本的实际计算可以只在几个时刻进行(这几个时刻由“控制单元” 1304确定)。 Generating timing adjustment as required, and the pulse position modulated LPOS signals (A and B can be encountered in pulse) matched filter required, the actual calculation of interpolated signal samples may be only a few moments (these time is determined by the "Control unit" 1304).

[0148] 假设在正常的伺服信道操作期间时序误差的值较小,那么就可以得到图14所示的时序基准产生系统1300的线性等价模型1400。 [0148] Assuming normal value timing error during track servo signal operation is small, so you can get the timing basis generation system shown in FIG. 14 1300 1400 linear equivalent model. 注意,这里所考虑的时序基准产生系统的模型很显著地不同于在通信接收器或者硬盘驱动器中通常遇到的离散时间误差跟踪同步器的模型,在后者中,在均勻隔开的时间间隔处计算时序误差估计,并且同步器需要运行的频率偏移量范围为标称频率的几个百分比的数量级。 Note that the model considered here the timing basis generation system is significantly different from the discrete-time communication receiver or a hard disk drive normally encountered error tracking synchronizers model, in the latter, at evenly spaced intervals calculated at the timing error estimate, a few percent of frequency offset range and synchronization needs to run for a nominal frequency of magnitude.

[0149] 参考图14,时序误差探测器1402的增益由双位脉冲在过零点处的导数的绝对值给出。 [0149] Referring to Figure 14, a timing error detector 1402 by double-digit gains in the past pulse gives the absolute value of the derivative at zero. 从(6)和(7)中,可以得到,所述增益正比于磁带速度,因此可以表示为KDv。 From (6) and (7) can be obtained, the gain is proportional to the tape speed, it can be expressed as KDv. 噪声样本ηη包括AWGN和插值步的非零长度所引入的量化噪声的贡献,如上所述。 Ηη AWGN noise samples including the contribution of non-zero length and interpolation step of quantization noise introduced, as described above. 相继时序误差估计之间的可变间隔通过包括一个开关1404来建模,所述开关在环滤波器1406的输入端在估计的过零点时刻(由指数η = ηζ(16)表征)提供插值信号,在其它时刻提供零误差信号。 Variable successive interval between the timing error estimate by including a switch 1404 is modeled, the switch in the loop filter input of 1406 the estimated zero crossing time (by the index η = ηζ (16) Characterization) provided an interpolation signal providing zero error signal at other times. 观察(17)和(18)可以看出,在将新的时序误差估计输入到环滤波器中并且计算新的时序校正项~„,之前,时序基准产生系统使用在过零点时刻所确定的标称步插值间隔的估计值7U。这个行为通过在环滤波器的支路中包括一个采样_保持元件1408来建模,该元件产生正比于时序误差估计的项。注意,保持间隔的可变长度与可变增益乘以时序误差估计的效果是等价的。为了补偿这个效应,需要引入一个项(在图14中用θη表示)来“均衡” 运用到每个时序误差估计的增益。由于过零点之间的最小距离为Ld = 5μπι,并且“控制单元” 1304具有有关上一次遇到的过零点和下一个过零点之间的距离信息,所以,独立于磁带速度的θ„的值由下式给出: L Observation (17) and (18) can be seen in the new timing error estimate is input to the loop filter and calculate a new timing correction term ~ "before, the timing basis generation system using zero-crossing time as determined by standard He said step interpolation interval estimate 7U. This behavior by the loop filter branch includes a sample holding member 1408 _ modeled, the element is proportional to the estimated timing error term. Note that variable length remains intervals variable gain multiplying effect of the timing error estimates are equivalent. To compensate for this effect, the need to introduce an entry (shown in Figure 14 by θη) to "balanced" use to estimate the gain for each timing error due to excessive The minimum distance between the zero point is Ld = 5μπι, and the "control unit" 1304 has been zero and the next one over the distance information between zero, so the tape speed is independent of θ on the last encounter, "the value represented by the following is given by: L

[0150] [0150]

Figure CN101361125BD00171

[0151] 其中,仏表示时亥时的过零点到下一个将要出现的过零点之间的距离(单位微米),ηζ为小于等于η的最大时间指数,在该时刻,时序误差估计值已经被输入到了环滤波器中,参见(18)。 [0151] where 仏 represents the maximum time when Haishi index over the next zero to the distance (in microns) between zero to appear too, ηζ less than or equal η, at which time, the timing error estimate has been entered to the loop filter, see (18).

[0152] 如前面所提到的,同步伺服信道400的时序基准产生系统1300必须在很宽的磁带速度范围内工作,而很宽的磁带速度范围又转换成很宽的相应的最小伺服信道信号带宽范围。 Timing reference [0152] As mentioned earlier, synchronous servo channel 400, generating system 1300 must operate over a wide tape speed range, and a wide tape speed range and converted into a wide corresponding minimum servo channel signal bandwidth range. 很清楚,这个事实对系统参数有不可忽略的影响,而系统参数的选择要使得时序基准产生系统的动态行为基本上保持独立于磁带速度。 Clearly, this fact can not be ignored on the system parameters influence the choice of system parameters such that the dynamic behavior of the timing basis generation system remains essentially independent of tape speed. 为了估计磁带速度变化时的系统性能,考虑图15所示的简化模型1500。 To estimate the performance of the system when the tape speed change, consider a simplified model 1500 shown in Figure 15. 如上所述,假设时序误差估计之间的可变间隔完全被可变增益θ „所补偿,并且假设磁带速度的估计值等于实际速度,BPv = ν,使得偏移项-I消失了。因此,图15中的环就等价于一个跟踪误差同步器(tracking error synchronizer),该同步器在输入端具有周期为Td = Ld/v微秒的循环平稳的(cyclostationary)信号。使用这个简化模型1500,时序基准产生系统1300的环带宽(其依赖于磁带速度)就由下式给出: [0153] As described above, assuming that the variable interval between the timing error estimate is completely variable gain θ "compensated, and assuming tape speed equal to the estimated value of the actual speed, BPv = ν, so that the bias term -I disappeared. Therefore, FIG. 15 is equivalent to a loop synchronizer tracking error (tracking error synchronizer), the synchronizer having at the input period Td = Ld / v microsecond Cyclostationarity (cyclostationary) signal. Using this simplified model 1500 the timing basis generation system 1300 loop bandwidth (which depends on the tape speed) on the following formula: [0153]

Figure CN101361125BD00172

[0154] 其中,闭环频率响应H(ζ ;ν)表示为: [0154] wherein the closed-loop frequency response H (ζ; ν) is expressed as:

[0155] [0155]

Figure CN101361125BD00173

[0156] 通过选择环参数 [0156] By selecting loop parameters

Figure CN101361125BD00174

,并且将H(z ;ν)和Td的表达式带入(23),可以得到,确定环的行为的积B(V)Td独立于磁带速度。 And the H (z; ν) and Td expressions into (23), can be obtained, determine the behavior of the product of the ring B (V) Td is independent of the tape speed. 例如,通过选择Kd= 1,y=l. IXlO-2, 和ζ =9.4\10_6,可以得到积8&)1^ = 0.18。 For example, by selecting a Kd = 1, y = l. IXlO-2, and ζ = 9.4 \ 10_6, the product can be obtained 8 &) 1 ^ = 0.18.

[0157] 时序基准产生系统1300的另一个要求是,在磁带加速和减速期间伺服信道能够可靠地工作。 [0157] Another requirement of the timing basis generation system 1300, in the tape during acceleration and deceleration servo channel work reliably. 注意,这个要求等价于要求在频率1/Td随时间线性变化的同时信道能够可靠地工作。 Note that this requirement is equivalent to the requirements of the channel can operate reliably at a frequency 1 / Td vary linearly with time at the same time. 众所周知,类似于图15中的简化模型1500的二阶时序恢复环在输入频率线性变化时会有非零的时序误差。 As we all know, a simplified model is similar to FIG. 15 1500 second-order timing recovery loop timing error in the input frequency linear change will be non-zero. 选择使用磁带速度估计值ί (参见(18))来周期性地更新所述步插值间隔估计值巧〜(该值由滤波器输出中的校正项来调整以确定插值时刻),避免了这个问题而不增加环的带宽或者环滤波器的阶数。 Choose to use the tape velocity estimates ί (see (18)) to update the step interpolation interval estimate clever - (the value from the output of the filter to adjust the correction term to determine the interpolation time), to avoid this problem periodically without increasing the bandwidth of the order of the loop or the loop filter. 所以,甚至在斜升斜降期间系统都可以可靠地运行。 Therefore, even during the ramp-up ramp system can reliably reduced. 于是,如果磁带速度是恒定的,那么在环滤波器的积分器中所累积的项就变得小得可以忽略。 Thus, if the tape speed is constant, then the loop filter integrator accumulated items becomes negligibly small. 注意,在这种情形中,可以在环滤波器之后乘以可变增益θ η,而不会显著地影响环的行为。 Note that, in this case, can be multiplied by the variable gain θ after the loop filter η, without significantly affecting the behavior of the ring. [0158] 通过仿真研究了时序基准产生系统1300的性能。 [0158] 1300 performance generating system timing reference through simulation studies. 图16A和16B显示了C脉冲的插值信号样本以及对于恒定的磁带速度(ν = 0. 5m/s(图16A)和ν = 12. 5m/s (图16B)) 的时序相位的收敛性。 16A and 16B show the interpolated signal samples C pulses and for constant tape velocity (ν = 0. 5m / s (Fig. 16A) and ν = 12. 5m / s (Fig. 16B)) timing phase of convergence. 在过零点时刻附近输入到环滤波器中作为时序误差估计值的样本用红色记号表示。 Zero crossing time input into the loop filter as close to the sample timing error estimates indicate red mark. 该结果是从用于产生伺服脉冲的PW50/2. 1 μ m = 0. 4的Lorentz型信道中获得的,AWGN产生的SNR = 25dB, ADC的采样频率fs = 15MHz,标称步插值距离Xi = 0. 25 μ m,速度估计值的初始误差为1 %,对于V = 0. 5m/s环滤波器参数等于^ /i = 1 x ICT1和^/ν = 2χ10"4,对于ν = 12. 5m/s 等于厂/ΐ = IxliT3禾口(/f = 2xl(T6。 The result is a servo pulses generated from PW50 for / 2. 1 μ m = Lorentz type 0.4 channel obtained, AWGN generated SNR = 25dB, ADC sampling frequency fs = 15MHz, the step interpolation distance Xi = 0. 25 μ m, the initial estimate of the error rate is 1%, for V = 0. 5m / s loop filter parameter is equal to ^ / i = 1 x ICT1 and ^ / ν = 2χ10 "4, for ν = 12 . 5m / s equal Factory / ΐ = IxliT3 Hekou (/ f = 2xl (T6.

[0159] 基于伺服信道输出信号的过零点产生估计值 [0159] Based on the servo channel output signal to produce an estimate of the zero crossings

[0160] 在前面的部分中认识到,同步伺服信道400的时序基准产生系统1300需要可靠的y位置和磁带速度估计值(分别用yest和^st表示)来确定所述步插值间隔的估计值和插值信号的过零点时刻的估计值以及各种环参数值。 [0160] In the previous section recognize that the timing reference 400 of synchronous servo channel generating system 1300 requires reliable y-position and tape velocity estimates (represented respectively yest and ^ st) to determine the estimated value of the step interpolation interval and the estimated value of the zero-crossing time and a variety of ring interpolated signal parameter values. 所以,产生yest和^st会带来问题,yest 和Vest的产生应该使用伺服信道ADC 406的输出中的信号样本来进行,使得yest和Vest的产生与时序基准的产生之间的耦合得以避免。 Therefore, to produce yest and ^ st cause problems, yest and Vest generated should use servo channel output signal samples in the ADC 406 to make the coupling between yest and Vest produce generation and the timing basis is avoided. 一个相关的问题是,对同步伺服信道操作的监视。 A related issue is that for synchronous servo channel operation monitoring. 必须设计一种监视系统,该系统能够早期探测到时序基准产生系统的偶然失谐(例如, 由伺服信道输出信号中的临时缺损来判断)并重新开始获取过程。 We must devise a monitoring system that can detect the early timing basis generation system accidentally detuning (for example, by a servo channel output signal to determine the temporary defect) and start the acquisition process. 另外,在这种情形中,必须通过直接观察伺服信道ADC的输出中的信号样本来获得一种解决办法。 In addition, in this case, it must be directly observed by the servo channel output signal samples in the ADC to obtain a solution.

[0161] 回想一下,在获取过程结束时磁带速度和y位置的估计分别由(3)和(4)给出,其中,在计算这些估计时考虑了在伺服信道ADC的输出中C、D和A脉冲的峰到达时间,仍然参见图6。 [0161] Recall that at the end of the acquisition process is estimated tape velocity and y positions are represented by (3) and (4), where, in the calculation of these estimates takes into account the output of the servo channel ADC in C, D and A pulse peak arrival times, still see Figure 6. 然而,基于峰的探测来产生y位置和磁带速度估计值会有下列缺点: However, detection of peaks generated based on y-position and tape velocity estimates have the following disadvantages:

[0162] a)双位峰的到达时间由一个运算来确定,该运算对信号的微分计算进行近似,所以会引入噪声增强。 [0162] a) double-digit peak arrival time is determined by an operation, the operation of the differential signal is calculated to approximate, it will introduce noise enhancement.

[0163] b)如果C、D和A脉冲中的一个或多个双位峰没有被探测到,那么就不可能可靠地产生y位置和磁带速度的估计值。 Estimate [0163] b) if C, D and A pulse of one or more two-bit peak has not been detected, it is impossible to reliably produce y-position and tape velocity. 为了处理探测中丢失峰的问题,插入虚设的峰到达时间来获得yest和u不是所希望的方法。 To deal with the problem of detecting the loss peak, insert dummy peak arrival time to get yest and u not a desirable approach.

[0164] c)探测到获取参数的缺失、恰当地重新开始产生y位置和磁带速度的估计值并使之生效会花很长的时间,在几个伺服帧的量级上。 [0164] c) to obtain the missing parameters detected and appropriate re-start generation y-position and tape velocity estimates and entry into force will take a long time, on the order of several servo frame.

[0165] 本发明还可以产生y位置和磁带速度估计值并同时监视同步伺服信道400的操作。 [0165] The present invention also can produce y-position and tape velocity estimates and simultaneously monitor the operation of synchronous servo channel 400. 该方法依赖于在伺服信道ADC的输出中确定信号的过零点以产生估计值yest和vest,以及依赖于观察同一信号的峰在时间间隔内的出现(这由有限状态机来确定)以监视同步伺服信道的操作。 The method relies on determining the signal at the output of the servo channel ADC's zero crossing to generate an estimated value yest and vest, as well as dependence on viewing the same signal peaks within the time interval (which is determined by a finite state machine) to monitor the synchronization servo channel operation. 使用时序基准产生系统的有限状态机“控制单元” 1304来提供所需要的观 The timing basis generation system using finite state machine "control unit" Concept 1304 to provide the required

察窗口。 Observation window.

[0166] 在初始参数获取过程期间,利用C、D和A脉冲中的相应双位之间的时间间隔的测量来确定磁带速度和y位置估计值。 [0166] During the initial parameter acquisition process, the use of C, D and measuring the time interval of pulses corresponding A two-position between the tape velocity and y to determine the position estimate. 为计算由(3)和(4)所给出的估计值,考虑双位的峰到达时间。 As calculated by (3) and (4) estimate given, considering the double-digit peak arrival time. 然而,在产生时序基准期间,关于伺服信道输出信号的时序相位的可靠信息通过观察伺服脉冲中的双位的过零点时刻来获得,而通过在匹配滤波之后确定伺服信号的峰的时刻而获得的时序信息受噪声增强的影响。 However, during the generation timing reference, reliable information on the timing of the phase servo channel output signal by observing the servo burst in the double-zero crossings of time to get, and by determining the time the peak of the servo signal after matched filtering obtained timing information by the noise enhancement effect. 这归因于这样的事实,即峰的时刻是由一种运算来确定的,该运算对信号微分的计算进行近似。 This is due to the fact that the peak time is determined by the kind of operation, the calculation of the approximate calculation of the differential signal. 所以,通过在(3)和(4)中引入时间间隔的测量(由伺服信道输出信号的过零点的时刻来获得),可以比使用受噪声增强影响的峰到达时间的测量能获得更可靠的估计值。 So, by introducing a time interval in (3) (4) and measured (by the zero-crossing moments servo channel output signal is obtained), the measurement time can be reached than by noise to enhance the impact of the peaks can be more reliable an estimated value. [0167] 这样,就避免了在探测中丢失了峰的情形中确定估计值yest和Vest的问题,因为时序基准产生系统1300提供有关观察间隔的信息,其中预期很可能伺服信道输出信号的过零点会出现。 [0167] Thus, it avoids the loss of the case peaks in the probe in determining the estimated value yest and Vest question, because the timing basis generation system 1300 to provide relevant observation interval information, which is expected to probably over zero servo channel output signal appears. 在观察间隔内没有过零点出现这种偶然情形中(例如,由于伺服信道输出信号中的缺损的缘故),使用由时序基准产生系统所提供的过零点的估计时刻。 In the observation interval no zero crossing occurs such occasional situations (for example, because the servo channel output signal defects sake), produce an estimate of time using the system provided by the zero crossing of the timing basis. 然而,如果缺损持续一段很长的时间,跨越几个伺服帧,那么,时序基准产生系统的失谐就会出现。 However, if the defect continue for a very long time, spanning several servo frame, then the timing basis generation detuning the system will appear. 为了探测这样一个事件,仍然监视由时序基准产生系统所确定的观察窗口内的伺服脉冲的双位峰的出现。 In order to detect such an event, still monitored by a system timing reference determined by observation window servo burst dibit peak production. 如果在每伺服帧所探测到的峰的数目小于某个固定阈值并且在预定时间间隔内一直保持低于该阈值,那么就宣告失锁(loss of lock)并重新开始初始获取过程。 If the number of peaks per servo frame detected is less than a fixed threshold value within a predetermined time interval, and remains below the threshold value, then declare loss of lock (loss of lock) and restart the initial acquisition process.

[0168] 如图4所示,y位置和磁带速度估计值的产生以及同步伺服信道操作的监视由两个功能块来进行。 [0168] As shown, produce y-position and tape velocity estimates and monitor synchronous servo channel operation consists of two functional blocks for 4. “异步监视功能块” 408确定过零点时刻并基于“控制单元” 1304所提供的观察间隔来探测双位峰的出现,“控制单元1304具有完整的与伺服帧的脉冲中的双位相关的时序信息,如前面所述。然后,“同步监视功能块”404计算y位置和磁带速度估计值并基于从“异步监视功能块” 408获得的信息监视时序基准产生功能块的操作。 "Asynchronous monitoring function block" 408 is determined based on the zero-crossing time and the "Control unit" 1304 provided observation interval to detect double-bit peak, "the control unit 1304 has a full pulse in the servo frame timing associated dibit information, as described above. Then, "synchronization monitoring function block" 404 computing y-position and tape velocity estimates based on information obtained from monitoring the timing "Asynchronous monitoring function block" 408 reference generating function block.

[0169] 图17显示了一个流程图,该图描述了异步监视功能块408的操作。 [0169] Figure 17 shows a flow chart which depicts the operation of asynchronous monitoring function block 408. 在获取过程结束时,下面的变量被初始化(步骤1700): At the end of the acquisition process, the following variables are initialized (step 1700):

[0170] a)acqFlag = 1,获取标记,表示同步伺服信道处于跟踪模式; [0170] a) acqFlag = 1, to obtain symbol indicates synchronous servo channel is in tracking mode;

[0171] b)k = 0,计数器,在每个采样时刻增加1并在每个伺服帧开始时复位; [0171] b) k = 0, counter incremented by one at each sampling instant and reset at the beginning of each servo frame;

[0172] c)k' = 0,索引号(index),在每个过零点时刻增加1并在每个伺服帧开始时复位; [0172] c) k '= 0, the index number (index), in each zero crossing time increased by 1 and reset at the beginning of each servo frame;

[0173] cDpeakDetFlagk' = 0,峰探测标记,在一个伺服帧内由过零点时刻所界定的第k' 个时间间隔开始时该标记置0,如果在该第k'个时间间隔内探测到正的双位峰,那么该标 [0173] cDpeakDetFlagk '= 0, peak detection mark, in a servo frame by the zero-crossing time as defined in section k' time interval is set to 0 at the beginning of the mark, if the detected within the 'time of the first interval to k n double-digit peak, then the standard

记置1。 Mind set.

[0174] 在每个采样时刻,“异步监视功能块”408首先检查“控制单元” 1304是否正在用信号表明出现了新的伺服帧(步骤1702)。 [0174] In each sampling time, "Asynchronous monitoring function block" 408 First check the "control unit" 1304 Are signals showed a new servo frame (step 1702). 回想一下,在一个伺服帧的末尾,即在一个B脉冲的最后一个双位的过零点处,对变量IntpCntReset进行断言以复位“计数器” 1306,将变量burstFlag置0以表明C脉冲的出现,也将变量dibitFlag置0以表明C脉冲中的第一个双位的出现。 Recall that at the end of a servo frame, that is a B pulse of the last two-bit zero crossing, asserts IntpCntReset variables to reset the "counter" 1306, the variable is set to 0 to signify burstFlag C pulses, also dibitFlag variable is set to 0 to indicate that C pulse in the first two-position appears. 所以,不管何时对IntpCntReset进行了断言,变量newFrame就置1。 Therefore, regardless of when to IntpCntReset been asserted variable newFrame on set. 在该情形中,计数器k和指数k'被复位(步骤1704)。 In this case, the counter index k and k 'is reset (step 1704). 在对新伺服帧的出现进行了检查之后,计数器k增加1(步骤1706)。 After the emergence of new servo frames were examined, the counter k is incremented by one (step 1706).

[0175] 然后,将计数器的值与第k'个观察间隔的上下限(wLk,,wRk.)进行比较(步骤1708),该上下限由“控制单元” 1304提供。 [0175] Then, the 'observations counter value of the k upper and lower limits (wLk ,, wRk.) Intervals (step 1708), the upper limit by the "control unit" provided 1304. 假设伺服信道信号的极性使得双位的正峰被首先探测到,那么,在第k'个观察间隔内预期探测到与伺服帧的第k'个双位相关的正峰以及过零点。 Suppose the polarity of the servo channel signal so dibit positive peak is first detected, then the first k 'within the observation interval is expected to detect the first servo frame k' a positive peak associated dibit and zero crossings. “控制单元” 1304也提供第三个变量(用wPk,表示),使得wLk, < wPk, < wRk'。 "Control unit" 1304 also provides a third variable (wPk, representation), making wLk, <wPk, <wRk '. 然后,(wLk,,wPk,)定义第k'个观察间隔中的子间隔,其中预期可以探测到所述双位峰,如图18所示的关于采样频率fs = 15MHz以及恒定磁带速度ν = 12. 5m/s的情况。 Then, 'observation interval (wLk ,, wPk,) in the definition of k-th sub-interval, wherein said expected to be detected dibit peaks, shown in Figure 18 on the sampling frequency fs = 15MHz and constant tape velocity ν = 12. 5m / s case. 如果wLk, <k<wPk,(步骤1710)并且信道输出样本rk超过了给定的阈值(用thres表示)(步骤1712),那么变量peakDetFlagk'置1 (步骤1714),以表示在第k'个时间间隔内探测到了双位峰。 If wLk, <k <wPk, (step 1710) and the channel output samples rk exceeds a given threshold (indicated by thres) (step 1712), the variable peakDetFlagk 'set (step 1714), as shown in the first k' a time interval to detect the double-peak. 如果wLk,<k<wRk,并且两个最新近的信道输出样本满足条件Iv1彡0和rk<0(步骤1716),那么就探测到了过零点并且在第k'个时间间隔中的过零点时刻计算为(步骤1718): If wLk, <k <wRk, and the two-channel output samples most recent satisfy the conditions Iv1 San 0 and rk <0 (step 1716), then to detect the zero crossing and at k 'time interval zero-crossing time calculated as (Step 1718):

[0176] [0176]

Figure CN101361125BD00201

[0177] 每个过零点时刻(在图18中用实心圆记号表示)都被记录为一个整数部分加一个小数部分,其中,整数部分由值k-Ι给出,而小数部分使用简单的查找表来确定。 [0177] Each zero crossing instant (shown in FIG. 18 by a solid circle mark) is recorded as an integer plus a fractional part of the section, which is given by the value of the integer part of k-Ι, and the fractional part using a simple lookup Table to determine.

[0178] 当k = WRk,时(步骤1720),观察间隔就期满了,指数k'增加1,并且变量PeakDetFlagk'被复位到0 (步骤1722)。 [0178] When k = WRk, (step 1720), on the expiration of the observation interval, the index k 'is incremented by one, and the variable PeakDetFlagk' is reset to 0 (step 1722). 如上所述,如果在观察间隔期满的时候没有探测到过零点,那么,变量tz,k,就取时序基准产生系统1300所提供的过零点的估计时刻值。 As described above, if the observation interval expires when there is no zero crossing is detected, then the variable tz, k, to produce an estimate of the time value provided by zero crossings of system 1300 taken timing reference.

[0179] 在图19中显示了“同步监视功能块” 1900的方框图。 [0179] shows a block diagram of a "synchronization monitoring function block" 1900 in FIG. 19. 当计数器1902达到观察窗口的上限时,即当k = wRk,时,变量PeakDetFlagk^ tz,k,的值就分别被输入累加器1904和延迟线1906。 1902 When the counter reaches the upper limit of the observation window, that is, when k = wRk, when variable PeakDetFlagk ^ tz, k, of the value is input accumulator 1904 and 1906, respectively, the delay line. 只要变量newFrame置1,即在每个伺服帧的开始处,计数器1902、累加器1904 和延迟线1906就被复位。 As long as the variable newFrame set, namely at the beginning of each servo frame counter 1902, accumulator 1904 and delay line 1906 is reset. 在A脉冲的末尾,可以对C、D和A脉冲中相应双位之间的时间间隔进行估计。 A pulse at the end, the time can be C, D, and A bursts dibit interval between corresponding estimate. 所以,在与A脉冲的最后一个双位相关的观察窗口的末尾将变量newEstimate 置1,并计算y位置和磁带速度估计值的新值。 So, at the end of last associated with a two-bit A pulse newEstimate observation window of the variable is set to 1, and calculate the new value of y-position and tape velocity estimates. 参考图6,C、D和A脉冲中的相应双位之间的时间间隔的测量由下式给出: With reference to FIG. 6, C, D, and A bursts in the respective measurement time interval between dibit is given by:

[0180] B1+B2+B3+B4 = sum「sum2 =、,。+、“+、,2+tz,3_ (tz,8+tz,9+tz,1(1+tz,n) (26) [0180] B1 + B2 + B3 + B4 = sum "sum2 = ,,. +," + ,, 2 + tz, 3_ (tz, 8 + tz, 9 + tz, 1 (1 + tz, n) (26 )

[0181] A1+A2+A3+A4 = sum3-sum2 = tz,4+tz,5+tz,6+tz,7-(tz,8+tz,9+tz,10+tz,n) (27) [0181] A1 + A2 + A3 + A4 = sum3-sum2 = tz, 4 + tz, 5 + tz, 6 + tz, 7- (tz, 8 + tz, 9 + tz, 10 + tz, n) (27 )

[0182] 然后,“同步监视功能块”404分别根据(3)和⑷计算出估计值yest和Vest的瞬时值,其中y。 [0182] Then, the "synchronization monitoring function block" 404, respectively, according to (3) and ⑷ calculate the instantaneous value of the estimated value yest and Vest, wherein y. . nst = l/[2tan(Ji/30)]和v。 nst = l / [2tan (Ji / 30)] and v. the . nst = 41fs。 nst = 41fs. 回想一下,磁带向前运动时1 = IOOym, Recall that when the tape moves forward 1 = IOOym,

而磁带反向运动时1 = 95μπι。 The reverse movement of the tape 1 = 95μπι. y位置和磁带速度估计值的平均值j)和i)由图19所示的电路1900中的一阶低通滤波器1908A、1908B来求出。 the average value of j y-position and tape velocity estimates) and i) circuit 1900 shown in Figure 19 first-order low-pass filter 1908A, 1908B obtains. 低通滤波器1908A、1908B的时间常数的选择是下面两个要求折衷的结果,即减小对估计值的瞬时值产生影响的噪声的要求和在计算估计量的平均值(由时序基准产生系统使用)中使等待时间与磁带驱动器中y位置和磁带速度的最大变化率的倒数相比要比较小的要求之间折衷的结果。 Low pass filter 1908A, 1908B time constant selection is the following two requirements compromise the result, i.e. to reduce the impact on the instantaneous value of the estimated value of the noise requirements and the amount of the estimated average calculation (by the timing basis generation system result of a compromise between the use of) manipulation compared with the reciprocal of the maximum waiting time rate of change of tape drives and tape speed y position to compare a small request.

[0183] 在每个帧的结尾,累加器1904中变量peakDetFlagk'的值表示已经探测的伺服脉冲中的双位的正峰的数目。 [0183] variable peakDetFlagk at the end of each frame, the accumulator 1904 'indicates the servo bursts have been detected in a number of positive peaks dibit. 然后,这个数与预定的阈值(用monThres表示)进行比较1910,并且使累加器1904复位。 Then, this number with a predetermined threshold value (indicated by monThres) comparing 1910 and 1904 so that the accumulator is reset. 如果对于相继的若干(等于预定值,用frameCntThres表示)伺服帧,在单个帧中探测到的峰的数目小于monThres,那么,就宣告失锁并重新开始获取过程。 If the number of successive (equal to a predetermined value, expressed frameCntThres) servo frame, the number of peaks detected in a single frame is less than monThres, then, declare loss of lock and re-start the acquisition process.

[0184] 通过仿真研究了用于产生y位置和磁带速度估计值的所述系统的性能。 [0184] The simulation study of the performance of the system used to generate the y-position and tape velocity estimates. 图20A、 20B和21A、21B分别显示了对于恒定的y位置值(y = Om)和恒定的磁带速度值(v = 6m/ s (图20A、21A)和ν = 12m/s (图20B、21B)) y位置估计的平均值和标准偏差以及磁带速度估计的标准偏差。 Figure 20A, 20B and 21A, 21B show the position for a constant value of y (y = Om) and constant tape velocity value (v = 6m / s (Fig. 20A, 21A), and ν = 12m / s (Fig. 20B, 21B)) y position estimate of the mean and standard deviation and the standard deviation of the tape velocity estimate. 没有显示磁带速度估计的平均值,因为平均值对实际值的偏差可以忽略。 It did not show the tape speed estimated average because the average deviation of the actual value is negligible. 在这两张图中,也显示了基于峰探测的系统的性能,以用于比较。 In both figures, also it shows the performance of the system based on peak detection, for comparison. 这些结果是从一个用于产生伺服脉冲的PW50/2. 1 μ m = 0. 4的Lorentz型信道中获得的,其ADC的采样频率fs = 15MHz。 These results are from PW50 for generating a servo pulse / 2. 1 μ m = Lorentz-type channel 0.4 is obtained, the sampling frequency of the ADC fs = 15MHz.

[0185] LPOS 符号(LPOS symbol)的最佳探测 [0185] LPOS symbol (LPOS symbol) Best Detection

[0186] 如前面所提到的,通过使A和B脉冲中的第二和第四双位的跃变偏移其标称的图形位置,如图1所示,伺服帧允许对LPOS信息进行编码而不影响y位置和磁带速度估计值的产生。 [0186] As previously mentioned, the A and B pulses through the second and fourth double-digit jump offset its nominal position graph shown in Figure 1, the servo frame allows LPOS information coding without impact y-position and tape velocity estimates. 注意,调制距离依赖于各磁带驱动器产品。 Note that the modulation distance depends on each tape drive products. 在由IBM所开发和销售的LTO产品(诸如3580型)中,调制距离为士0.25 μ m,而在IBM企业产品(诸如3592型)中,调制距离为士0.5μπι。 By IBM LTO developed and marketed products (such as a 3580 model), the modulation distance is 0.25 μ m, while in the IBM enterprise products (such as a 3592 model), the modulation distance disabilities 0.5μπι. 在许多异步伺服信道中,LPOS信息的探测基于在伺服信道输出中对双位信号样本的峰的偏移的观察。 In many asynchronous servo channels in the servo channel output signal sample of dibit peak of LPOS offset information based on observation probe. 对于如图22Α所示的等于1的编码了的LPOS符号,A和B脉冲中双位之间的8个间隔的测量用a到h来标记。 For 22Α shown in Figure 1 is equal to the coding of LPOS symbols, eight double-spaced between measurement with a to h to mark the A and B pulses. 探测器根据图22B中的表格对相应的各间隔的测量进行比较,并应用大数判决解码规则(majority decoding rule)来确定编码了的LPOS符号,其中该规则要求四个可能条件中至少三个条件为真。 At least three detectors in accordance with the table of FIG. 22B each corresponding measurement interval are compared, and the application of large numbers decision decoding rules (majority decoding rule) to determine the encoded LPOS symbol, wherein the rule requires four possible conditions the condition is true.

[0187] 基于峰的探测以及峰到达时间的记录的常规方法有下列局限: [0187] peak detection and peak arrival conventional methods based on the recording time has the following limitations:

[0188] a)基于对间隔的测量(通过求峰的到达时间之差来测量)的大数判决解码规则不是LPOS符号的最佳探测方案,其中使用脉冲位置调制(puis印osition modulation, PPM) 技术对LPOS符号进行了编码。 [0188] a) based on the measurement of the interval (measured by the difference between the arrival time of peak demand by) the largest number of decision rules are not optimal detection scheme decoded LPOS symbols, which use pulse position modulation (puis India osition modulation, PPM) technology LPOS symbols are encoded.

[0189] b)如果在图22中的表格所列出的用于符号判决的四个条件中有两个条件被满足,除了诉诸抛硬币的办法,不清楚该如何处理。 [0189] b) If there are two conditions are met in the four conditions for the decision in the symbol table listed in Figure 22, in addition to resort to a coin flip way, I do not know how to deal with.

[0190] c)不可能有与LPOS符号判决相关的可靠性的度量。 [0190] c) there can be no measure of reliability associated with LPOS symbol decisions.

[0191] 本发明还支持LPOS符号的最佳探测以及对LPOS探测过程的可靠性的同步监视。 [0191] The present invention also supports optimum LPOS symbol detection and monitoring of the reliability of synchronous LPOS detection process. 这个方法以度量(the metric)为基础,对于每个LPOS符号Q1G {-1,+1},所述度量给出两个假设二-1和^^ =1的可能性。 This approach to measure (the metric) basis, for each LPOS symbol Q1G {-1, + 1}, the metric is given two assumptions ^^ = - 1 and b 1 of possibilities. 通过使用假设检验所采用的度量,可以引入与LPOS探测相关的信号_噪声加失真之比的量度,这就允许监视LPOS探测过程以及监视各个LPOS 符号判决的可靠性。 By using hypothesis testing metrics employed, a measure can be introduced to detect signals associated with LPOS _ noise plus distortion ratio, which allows monitoring LPOS detection process and monitor the reliability of each LPOS symbol decisions. 具体说,因为在磁带系统中总是有两个专用伺服信道,从中可以得到LPOS信息,所以,这个新引入的可靠性量度可以容易地用来判断所述两个信道中的哪个信道提供最可靠的LPOS符号判决。 Specifically, since the tape systems there are always two dedicated servo channels from which LPOS information can be obtained, so that the newly introduced reliability measure may be readily used to determine which of the two channels provides the most reliable channel The LPOS symbol decisions.

[0192] 回想一下双位信号脉冲的表达式(7)和伺服信道输出信号的表达式(8)和(9), 以及用于在AWGN存在时从给定的设置中探测波形信号的最佳接收器的公式。 [0192] Recall that the expression dibit signal pulse (7) and expression servo channel output signal (8) and (9), and for the best when AWGN presence from a given set of waveform signal detected Receiver formula. 观察(8)和(9)可以看到,通过将脉冲位置调制运用到A和B脉冲的第二和第四双位可以实现对LPOS 符号的编码。 Watch (8) and (9) can be seen by the pulse position modulation applied to the A and B pulses of the second and fourth dual bit can be achieved on LPOS symbol coding. 所以,对于恒定的磁带速度,与假设民,=,α e {-1,+1}的可能性相关的度量可以表示为: Therefore, for a constant tape speed, and assuming people, = , α e {-1, + 1} associated probability metric can be expressed as:

[0193] (H …=α)=〜, [0193] (H ... = α) = ~,

Tf ^ y 3T„ 2 vtai(π /30) 2 Tf ^ y 3T "2 vtai (π / 30) 2

ίτ 丨丨yt T<t ίτ Shushu yt T <t

F 2 vtan(lt/30) 2 F 2 vtan (lt / 30) 2

M M

άίλ- άίλ-

[0194] [0194]

[0195] [0195]

iT Jf xy JTd F 2 ' vtao(π /30) 2 We Jf xy JTd F 2 'vtao (π / 30) 2

ITr ITr

STj STj

2 2

Λ + Λ +

ί2>+ ί2> +

ν tan (π/30) 37>__ y ν tan (π / 30) 37> __ y

iTF iTF

ITp ITp

^_■玛 ^ _ ■ Mary

(tc/30) 2 (Tc / 30) 2

i;tBD(π /30) 2 i; tBD (π / 30) 2

dt+ dt +

21[0196] 21 [0196]

Figure CN101361125BD00221

[0197] 注意,由(28)所给出的度量等价于由匹配滤波器接收器所获得的度量。 [0197] Note that measures from (28) is equivalent to the measure given by the matched filter receiver is obtained. 回想一下对应关系1^= {0,1},引入变量的变化χ = vt,并定义匹配滤波器波形s (x;y)(其中β =0,1,j = 1,2)为 Recall correspondence between 1 ^ = {0,1}, the introduction of changes in the variables χ = vt, and the definition of a matched filter waveform s (x; y) (where β = 0,1, j = 1,2) to

[0198] [0198]

Figure CN101361125BD00222

[0199]则度量的表达式(28)变为 [0199] to measure the expression (28) becomes

Figure CN101361125BD00223

[0204] 注意,度量的表达式(30)不依赖于磁带速度。 [0204] Note that measure the expression (30) does not depend on the speed of the tape. 回想一下,时序基准产生系统提供时刻序列{tn},使得可以以固定的频率(每微米1/Xi个样本)来获得插值信号样本Ir (tn)},而不依赖于磁带速度。 Recall that the timing basis generation system provides the time sequence {tn}, so that the frequency can be fixed (per micrometer 1 / Xi samples) to obtain interpolated signal samples Ir (tn)}, without depending on the tape speed. 所以,假设采样频率足够大从而能够避免混叠效应,并假设Nf = Vxi为整数,那么,可以使用插值信号样本在数字域中计算所述度量: Therefore, assuming that the sampling frequency is large enough to be able to avoid aliasing effects, and assuming Nf = Vxi is an integer, then you can use the interpolated signal samples in the digital domain calculating the measure:

[0205] [0205]

Figure CN101361125BD00224

[0207]其中3,Xy),i = 1,2,3,4,表示整数集合,其定义为 [0207] wherein 3, Xy), i = 1,2,3,4, denote the set of integers, which is defined as

Figure CN101361125BD00231

[0212] 本发明中的最佳LPOS符号探测器2300计算得到值m^,β =0,1,并对其进行比较以产生对在第1个伺服帧中编码的LPOS符号的判决&,如图23所示,即: [0212] The present invention is best LPOS symbol detector 2300 calculated value m ^, β = 0,1, and compare them to produce encoded at a servo frame LPOS symbol decision & such 23, namely:

[0213] [0213]

Figure CN101361125BD00232

[0214]波形 s广_)(x ;y) (β = 0,1)示于图24,其中PW50 = 0. 84μπι,y = 0,以及Xi = 0. 25 μ m。 [0214] Waveform s wide _) (x; y) (β = 0,1) is shown in Figure 24, where PW50 = 0. 84μπι, y = 0, and Xi = 0. 25 μ m.

[0215] 考虑使用具有由集合3,00,i = 1,2,3,4所指定的指数的子集来进行(31)中的求和,可以显著地减小度量计算所要求的复杂性。 [0215] consider having a collection of 3,00, i = 1,2,3,4 specified subset of the index to be (31) summing measure can significantly reduce the complexity of the calculations required . 具体说,在每个集合中只考虑4个指数,这4个指数对应着波形8{^)0^;》(0 =0,1, j = 1,2)的最大绝对值,这样就可以获得度量(31)的近似,该近似能提供可靠的LPOS符号判决。 Specifically, considering only four index in each set, which corresponds to the four index waveform 8 {^) 0 ^; "(0 = 0,1, j = 1,2) maximum absolute value, so that you can obtain metric (31) approximation, this approximation provides reliable LPOS symbol decisions. 于是,在度量计算中考虑使用的(力, i = 1,2,3,4的指数的子集就由下式给出: Thus, in considering the use of metric calculation (force, i = 1,2,3,4 subset index on is given by:

[0216] [0216]

Figure CN101361125BD00233

[0220] 其中,Lp = 2. 1 μ m表示一个双位的正峰和负峰之间的距离。 Distance [0220] where, Lp = 2. 1 μ m represents a double-digit positive and negative peaks between. 例如,在图24A和24B 中用较大的标记来表示在度量(31)的近似计算中所考虑的波形〜ω(χ ;y) (β = 0,1)的样本。 For example, FIG. 24A and 24B with a relatively large waveform ~ω marked to indicate the measure (31) considered in the approximation (χ; y) (β = 0,1) of the sample.

[0221] 图25显示了简化的LPOS符号探测系统2500的方框图。 [0221] Figure 25 shows a block diagram of a simplified LPOS symbol detection system 2500. 如前面所述,“控制单元” 1304具有关于伺服帧中的脉冲中的双位的完整的时序信息,由该“控制单元” 1304提供每帧16个插值时刻的值ItJ,在这些时刻估计对度量产生贡献的项,参见(34)。 As mentioned earlier, "control unit" 1304 has complete information about the servo frame timing pulses in the two-position, providing 16 per frame interpolation time value ItJ by this "control unit" 1304, in these moments estimate measure the contribution of an item produced, see (34). 有关所述时刻{tn}的信息也允许恰当地选择匹配滤波器波形的样本,如图25中的方框图所示。 Information about the timing {tn} also allows the proper selection of the matched filter waveform samples, a block diagram is shown in Figure 25. 在由“控制单元” 1304所确定的每个时刻tn处,计算插值信号样本,从中减去所选择的波形样 In the "control unit" by the 1304 tn determined at each moment, the interpolated signal samples, subtract the selected waveform samples

=0,1)。 = 0,1). 对所得到的差求平方并累加起来以形成两个度量值me, β = 0,1。 The resulting difference of squaring and add up to form two metric me, β = 0,1. 当在求和间隔的末尾(出现在B脉冲的第四个双位的末尾)“控制单元” 1304将变量newSymbolDecision置1时,根据规则(33)就探测到一个新的LPOS符号,并使累加器复位。 When (the end of the fourth double-digit appeared in B pulse) at the end of the summation interval "control unit" 1304 newSymbolDecision variable set to 1, according to the rules (33) to detect a new LPOS symbols, and to accumulate reset. 在探测系统的输出中也呈现出了低度量值和高度量值(分别用Mcl和Mu示,并分别指示在正确和不正确的假设下的度量),以供进一步的处理来确定LPOS探测过程的可靠性测度。 The output of the detection system also showed a low value and high value (respectively Mcl and Mu show, and indicate measure in correct and incorrect assumptions), for further processing to determine the LPOS detection process The reliability measures.

[0222] 度量的表达式(31)是在恒定磁带速度的假设下导出的。 [0222] measure the expression (31) is under the assumption of a constant tape speed export. 然而,如前面所说明的, 速度估计值被输入时序基准产生系统1300来跟踪随时间变化的速度。 However, as previously explained, the velocity estimation value is input to the timing basis generation system 1300 to track time-varying velocity. 所以,即使在磁带加速运动和减速运动期间也可以获得可靠的LPOS符号判决。 So, even during tape acceleration and deceleration to be reliable LPOS symbol decisions. 图26A和26B显示了特征为SNR 等于25dB的AWGN伺服信道的仿真结果,该结果给出了磁带加速期间的估计速度(图26A) 和计算出的度量值(图26B),其中初始磁带速度等于0. 5m/s,磁带加速度等于lOm/s2。 26A and 26B show characteristics equal to 25dB SNR of AWGN servo channel simulation results that gives the estimated speed of the tape during acceleration (FIG. 26A) and the calculated metric (Fig. 26B), wherein the initial tape speed is equal to 0. 5m / s, the tape acceleration equal lOm / s2.

[0223] 为了确定探测过程的质量,定义探测点处的平均信号_噪声加失真之比。 [0223] In order to determine the quality of the detection process, the average signal _ defined probe points at the noise plus distortion ratio. 引入量Mc和〃^来分别表示在正确的假设下所得到的度量的平均值和方差,引入量M1和σΙ,来分别表示在不正确的假设下所得到的度量的平均值和方差,于是,探测点处的平均信号-噪声加失真之比被定义为: 〃 ^ introduction amount Mc and to represent in the correct assumption that the resulting measure of the mean and variance, introduced the M1 and σΙ, represented in the incorrect assumption that the resulting measure of the mean and variance, respectively, so The average signal detection point - more than the noise plus distortion is defined as:

f \ \Mr - McI f \ \ Mr - McI

[0224] SDRestj : 201og g 2'. [0224] SDRestj: 201og g 2 '.

(W^cJ (35) (W ^ cJ (35)

[0225] 与LPOS符号判决相关的可靠性的测度由下式给出: [0225] Measurement and LPOS symbol decisions concerning reliability is given by:

Ίμ μ Π Ίμ μ Π

[0226] RELest f =2Q\og ' , 7^ ~ 2C/' (36) [0226] RELest f = 2Q \ og ', 7 ^ ~ 2C /' (36)

[0227] 图27显示了用于产生平均信号-噪声加失真之比的估计值和产生每个LPOS符号判决的可靠性的估计值的系统2700的方框图。 [0227] Figure 27 shows the average signal for generating - a block diagram of the system reliability of the estimated value of the estimated value of the noise plus distortion ratio and generating each LPOS symbol decisions in 2700. 如上所述,SDResta和RELestJ可以分别用来监视LPOS探测过程以及从并行工作的两个伺服信道所探测到的LPOS符号中选择最可靠的。 As described above, SDResta and RELestJ can be used to monitor each LPOS detection process and from work in parallel two servo channels are detected LPOS symbols to select the most reliable.

[0228] 对基于同步伺服信道的原型LPOS符号探测器的性能进行了测量,并与当前用在LTO磁带驱动器产品中的基于峰探测的常规LPOS符号探测器的性能进行了比较。 [0228] The prototype LPOS symbol detector based on the synchronous servo channel properties were measured and compared with currently used in LTO tape drive products in the peak detection routine LPOS symbol detector based performance were compared. 利用FPGA实现了同步伺服信道和TBS系统,并使用磁带驱动器伺服信道ADC的输出信号作为这两个系统的输入来并行地运行了所述同步伺服信道和TBS系统。 Using FPGA to achieve a synchronous servo channel and the TBS system, and using a tape drive servo channel ADC output signal as an input to the two systems run in parallel synchronous servo channel and the TBS system. 使用相同的LPOS词解码器(LPOS word decoders)对LPOS符号序列进行解码并对两个信道中的判决误差数进行测量,不使用误差校正。 Using the same LPOS word decoders (LPOS word decoders) sequence of LPOS symbol decision error and the number of decoding two channel measurement, error correction is not used. 利用伺服读磁头在磁带上的12个不同的横向位置处(用卷数(wrap numbers)表示)对通常在正常的磁带驱动器运行期间提供伺服信息的两个物理伺服信道CHA和CHB进行读(磁带速度等于6. 22m/s)。 Use 12 different positions lateral servo read head on the tape (indicated by the number of volumes (wrap numbers)) of typically provide servo information during normal tape drive runs two physical servo channels CHA and CHB read (Tape speed is equal to 6. 22m / s). 通过每卷(wrap)读取80000个LPOS词获得了LPOS词误差(word errors)数的测量,并对每卷都重复这种测量。 By reading 80000 LPOS words per roll (wrap) won the LPOS word errors (word errors) measuring the number, and each volume repeated this measurement. 结果示于表I。 The results are shown in Table I.

[0229] 表I [0229] Table I

[0230] TBS和同步伺服系统之间的性能比较 [0230] Performance Comparison between TBS and synchronous servo system

[0231] [0231]

Figure CN101361125BD00251

[0232] 注意,误差总数不包括从卷#30和卷#46所获得的结果,因为在这些情形中,同步误差妨碍了TBS系统的正常运行。 [0232] Note that the total number of error does not include the results obtained from the volume # 30 and volume # 46, because in these cases, the synchronization error impede the normal operation of the TBS system. 具体说,TBS系统的同步误差是由于间隙探测器(gap detector)在伺服带边缘的失效所致。 Specifically, TBS system synchronization error is due to a gap detector (gap detector) due to the failure of the servo band edge. 在正常工作条件下两个系统的总误差数之间的比较表明,基于同步伺服信道体系结构的LPOS探测系统的误差率比基于TBS的LPOS探测系统所显示的误差率好约47倍。 Under normal operating conditions are the total number of errors between the two systems showed an error rate of about 47 times based on synchronous servo channel architecture LPOS detection system is better than the error rate of LPOS detection system based TBS displayed. 尽管表I所示的结果是由某个具体磁带驱动器和某个具体磁带盒所获得的,但预期这两种系统之间的性能差异一般不会与这里所示的差异有本质上的不同。 Although the results in Table I illustrated by a specific tape drive and a specific cassette obtained, but the difference in performance between the two systems is generally not expected and shown here has a different nature.

[0233] 需要注意的是,尽管本发明是在完全发挥作用的数据处理系统中进行描述的,但那些在本领域中具有普通技术的人员会认识到,本发明中的过程能够以指令的计算机可读介质的形式以及各种其它形式进行分发,并且不管实际上用来进行分发的负载信号的介质是什么具体类型,本发明都适用。 [0233] It should be noted that although the present invention has been fully functioning data processing system has been described, but those who are of ordinary skill in the art will recognize that the process of the present invention can command the computer readable form of media, and various other forms of distribution, and regardless of the medium used for actually distributing load signal what specific type of the present invention are applicable. 计算机可读介质的例子包括可记录类型的介质,诸如软盘、硬盘驱动器、RAM、CD-ROM,以及发送类型的介质,诸如数字和模拟通信链接。 Examples of computer readable media include recordable-type media, such as floppy disks, hard drives, RAM, CD-ROM, and transmission type media such as digital and analog communication links.

[0234] 本发明的描述只是用来进行说明和描述,并不打算是详尽无遗的或限制在这里所给出的形式中。 [0234] The invention is described only used for illustration and description, not intended to be exhaustive or to limit the form given here. 对那些在本领域中具有普通技术的人员来说,很显然可以进行许多修正和改变。 For those who have ordinary skill in the art it is readily apparent that many modifications and changes may be. 实施例的选择和描述是为了最好地说明本发明的原理、实际应用,以及为了使本领域中具有普通技术的其它人员对于适合于所能想象到的具体应用的具有各种修正的各种实施例能够理解本发明。 Embodiments were chosen and described in order to best explain the principles of the present invention, the practical application, and to enable others skilled in the art having ordinary skill suited for a particular application can imagine having a variety of modified variety embodiments of the present invention can be understood. 此外,尽管上面针对方法和系统进行了描述,但本领域中的需求也可以通过计算机程序产品来满足,其中这些计算机程序产品中包含了用于在磁带存储系统中进行纵向位置(LPOS)探测的指令。 Moreover, although described above with respect to methods and systems have been described, those skilled in the requirements to be satisfied by a computer program product, wherein the product comprises a computer program for performing the longitudinal position of the tape storage system (LPOS) detection instructions.

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
CN1283851A2 Aug 200014 Feb 2001国际商业机器公司Tough detection to servo datas modulated as time base
CN1604196A3 Aug 20046 Apr 2005国际商业机器公司Tape servo information with superimposed data information providing servo band identification
US2003/0016465 Title not available
Classifications
International ClassificationG11B5/584
Cooperative ClassificationG11B5/584
European ClassificationG11B5/584
Legal Events
DateCodeEventDescription
4 Feb 2009C06Publication
1 Apr 2009C10Request of examination as to substance
12 Jan 2011C14Granted