FR2779863A1 - Longitudinal track following system for multitrack tape, especially for magnetic and optical recordings - Google Patents

Abstract

A charge coupled device (CCD) type detector having typically 96 elements reads samples from 80 tracks on a tape (BD). The detector signals are fed to a processor (CC) via an analogue/digital convertor (CAN), and the results are used in a controller of the relative actuator position (DTR). The controller determines, from neighboring tracks, the coefficients to be applied to a main track which after integration furnish a detector correction signal.

Description

TRACK TRACKING SYSTEM FOR RECORDING / PLAYBACK

  OF AN INFORMATION MEDIUM AND RECORDING MEDIUM

  The invention relates to a tracking system for recording / reading an information medium, in particular for a multitrack information medium such as a multitrack magnetic tape.

  arranged longitudinally on the strip.

  The invention applies in particular to the reading of magnetic or optical recordings and, in this context, to the reading of high density recordings. It finds a preferential application in recording systems such as computer peripherals and all professional systems. It can be extended to recordings on optical tape and on magnetic or optical disc, if you want to read several

parallel tracks.

  High density recording on parallel tracks poses a double problem for track replay and track separation. The small width of the tracks (less than 20 μm) makes it difficult on a tape reading to ensure the accuracy of the tracking on the basis of the only guidance

band edge mechanics.

  The need to guarantee the interoperability of bands and

  readers compound this difficulty.

  Furthermore, obtaining a good signal-to-noise ratio in reading requires re-reading the entire track width, which excludes the existence of a guard between tracks and induces crosstalk phenomena of reading

track to track.

  The increase in the longitudinal track density of the linear recording / playback systems makes it necessary to install efficient track tracking systems, making it possible to position the reading system in front of the written tracks with a small residual error in front of the track. track width. The positioning and guiding precision given by the conveyor belt mechanics is no longer

  sufficient to naturally guarantee good positioning.

  Some systems allow recordings on contiguous tracks with a pitch of less than 10 µm. Reading of such a recording can be done optically as described in French patent n 89 17313 (see FIG. 1). A TL6 multi-pixel detection device receives a light beam carrying information read from the magnetic strip BD. So that the pixels receive the track information from the strip and that everyone can read a determined track, a TL5 system for deflecting the read beam is provided between the strip and the detection device. For example, a mirror or blade activator allows the deflection in reading optical beam at a sufficient speed to permanently ensure the correct positioning of the pixels of the detection device by

compared to the corresponding tracks.

  The tracking systems usually implemented in linear recorders are based on the permanent or periodic reading of one or more control tracks on which defined signals ("pilot" frequencies for example) have been recorded. A position error signal is extracted from the signal processing of this (these) track (s), and acts on an actuator, which performs a relative displacement of the beam of

  reading against detectors.

  In the case of multitrack technology with optical reading, simultaneous access to a large number of contiguous tracks makes it possible to directly develop an error signal from the observation of the crosstalk of reading between these tracks. The error signal can then be used to

  check the position of the actuator.

  French Patent No. 92 15474 describes a servo device based on this process. This system ensures a "relative" positioning, that is to say it allows the correct centering of the photosensitive elements of the reading detector (pixels) on the center of the replayed tracks. It is completed by a second device which generates an error signal if the relative positioning does not match the pixel j to the

track j.

  The invention provides a system which has a shorter response time than known systems and which allows faster correction of tracking. In addition, this system is less restrictive for the recording format. The invention therefore relates to a system for reading a magnetic medium having several tracks of information readable in parallel, and comprising a detection device comprising at least as many detectors as there are tracks, making it possible to read simultaneously and at regular intervals a sample of information on each track, said detection device comprising a parallel / serial shift register receiving in parallel the samples of information read by the detectors at each instant of reading and retransmitting them in serial form, characterized in that that it comprises: * a processing circuit receiving each sample of information to be processed from each track as well as the sample from a first neighboring track and the sample from a second neighboring track, multiplying the value of said sample: - by +1 when the sample of the first neighboring track is negative and the sample of the second neighboring track is po sitive; - par-1 when the sample of the first neighboring track is positive and the sample of the second neighboring track is negative; - by 0 when the samples of the neighboring tracks have the same sign; To an integration circuit receiving each sample value thus multiplied, integrating said values obtained at each reading instant, then integrating the values obtained at the following reading instants; * a relative tracking control circuit receiving the integration result of the integrating circuit and providing a tracking control signal from the detection device. The invention also relates to a recording medium comprising several tracks recordable in parallel, characterized in that said tracks each comprise a preamble area recorded or recordable in parallel, said areas containing information

  to locate the moon tracks in relation to the others.

  The various objects and characteristics of the invention will appear

  more clearly in the description which follows, given by way of example and

  in the appended figures which represent: - Figure 1, a magnetic tape reading system known in the art; - Figure 2, a general embodiment of the system of the invention; - Figure 3, an example of positioning detectors relative to the tracks of a magnetic strip; - Figure 4, a general embodiment of a circuit for detecting the relative position of the tracks of a recording medium with respect to a detection device; - Figure 5, a detailed embodiment of the circuit of Figure 4; - Figure 6, a detection device associated with a multitrack recording medium; - Figure 7, an example of organization of an absolute tracking system according to the invention; - Figure 8, an example of a circuit for implementing an absolute tracking control circuit according to the invention. According to the invention, the function for generating the error signal is independent of the crosstalk correction system, which allows independent optimization of the two functions. Furthermore, the absolute positioning method no longer requires using a particular modulation code to which violations have just been made, but is based on zones independent of the data zones, thus making it possible to use any type of modulation code adapted to the channel. The particular signals used for absolute positioning can be chosen according to the

  desired overall performance (absolute position detection time).

  Finally, absolute and relative positioning are now dissociated, which

  allows a much shorter establishment time.

  The general operation is as follows: - the reading system has a greater number of elements (pixels) than the number of tracks; - the relative positioning loop ensures alignment of the pixels on the nearest tracks; - an absolute position detection system determines where the physical tracks are located, or, which is equivalent, indicates the pixel number corresponding to the first track (track 0); The reading electronics process the signals of all the pixels; Track position information is transmitted to the decoded data processing electronics, which

  only processes data from physical tracks.

  FIG. 2 represents a general diagram of the system of the invention. This system includes a TL6 detection device comprising a row of tli detectors. The width of this row of detectors is wider than the width of the strip BD to be read. For example, as shown in FIG. 2, for a number of tracks of 80, the detector device TL6 comprises 96 detectors tli. The detection device TL6 comprises, in a known manner, a CCD circuit, not shown, making it possible to transmit the signals detected on the tracks to processing circuits. These processing circuits include an analog-digital converter ADC, a digital processor CC capable of processing the information of the 96 detectors. A relative positioning circuit DTR of the detection device with respect to the tracks is associated with the processor CC and controls a control system TL5 of the position of the detection device with respect to the tracks. A circuit for detecting the absolute position DTA of the detectors relative to the tracks is also associated with the processor for

  indicate the positional relation of the detectors in relation to the tracks.

  We will first describe the design and an example of

  realization of the DTR relative positioning circuit.

  In an ideal multitrack reading system, each reading element i delivers at time k a signal xi (k) proportional to the symbol ai (k)

written on this track.

  In practice, the read signal undergoes deformations due to 1- the presence of noise b (k) 2- interference between successive symbols of the same track (intersymbol interference) due to the limited bandwidth of the channel 3- interference between symbols of neighboring tracks (crosstalk between tracks) as a result of faults in separation between tracks of the playback system 4- interference between symbols of neighboring tracks due to faults in positioning and guiding the support

  recording or writing / reading heads.

  FIG. 3 represents a set of parallel tracks of width w, and the reading system, positioned with a deviation d from

the optimal position.

  The read signal of the head element i at the instant k is equal to: Xi (k) = (1-) ai (k) + w ai-l (k) + b (k)

  Suppose we know the signal ai-l (k) written on track i-

  1. It is possible to calculate the intercorrelation of Xi (k) and ai-l (k).

  C..1 = k Xi (k) ai_1 (k) i d d C _1 =. N [(1- w) a i ± 1 (k) + b (k)] ai 1 (k) 1 d) d C = - [(1 ±). a) 'ai -lk-.a. a + i_ ii1 N (wi (k) i-1 (k) + N w i-1 (k) i-1 (k) N ak) The first term represents the intercorrelation of the signals written on tracks i and i -1. If the track signals are statistically

independent, this term is null.

  Likewise, the last term represents the intercorrelation of noise and

  of the signal. If the noise is independent of the signal, this term is zero.

  The second term represents the autocorrelation of the signal from track i-1. It is the signal energy over n samples. It’s constant on average,

  and we will take it equal to 1 by convention.

  Then I d d C. 1 E -a a Cii-1: Nw ai-1 (k) ai-1 (k) w In the case shown in the figure, Ci, i + 1 is zero. If the deviation of the detector from the tracks is in the other direction, Cii_1 = 0 d Cii + 1 w We can therefore extract a position error signal based on the signals of track i: CC d Cii_1 -i + 1 e -W In order to improve the quality of this signal, it is advantageous to make this calculation for each of the tracks, and to average the results dl1 e [Ci_ - Ci, i + 1] According to the invention, performs the calculation directly, in order to overcome the constraints of response time and dimensioned rounding

  for optimum operation of the crosstalk corrector.

  At time k: F (k): lC iJ C (k) - Ci 'i + l (k) i Cii-1 = N xi (k) ai-1 (k) The signals ai (k) recorded on the tracks are approximated by the sign of the samples read over on this track. This simplifies decoding of the signal. The advantage of this simplified decoding is that it is not necessary to wait for the convergence of the various signal processing devices that are generally found in the read channel (adaptive equalizer, PLL ...) to have of the decoded signal. The approximation only generates additional noise in the estimation of the position error: C .. # 1g () Ci1 N xi (k). sgn (xi_1 (k)) where: s (k) = N i k [xi (k). sgn (xil 1 (k)) - xi (k). sgn (xi + l1 (k))] NP i k or again: s (k) = - NP = xi (k). [sgn (xi_1 (k)) - sgn (xi + 1 (k))] NP i k 1 This last form allows an implementation of the algorithm

particularly simple.

  FIG. 4 represents a simplified embodiment example of the

  DTR relative positioning circuit.

  This circuit receives on its input IN the samples of information transmitted on the serial output of the CCD of the detection device TL6. Each successive sample is transmitted to a multiplier M1 where it is multiplied by a coefficient 1, -1 or 0 provided by the DS circuit according to the signs of

neighboring samples.

  The samples thus multiplied are integrated into an integration circuit II. Integrator output OUT provides an error signal to correct the relative position error of the detection device

compared to the slopes.

  The circuit DS makes it possible to determine the multiplication coefficient K of crosstalk correction for each sample. For a sample xi, this coefficient K take one of the values -1, +1 or 0 by applying the following table: Sign value of Sign of coefficient K the sample xi-, the sample xi, +

-1> 0 <0

+1 <0> 0

0> 0> 0

0 <0 <0

  We therefore see that if the samples have the same sign the coefficient K is zero. If they have opposite signs, the coefficient K has for

  value +1 or -1 as indicated in this table.

  FIG. 5 represents a detailed embodiment of the DTR circuit. This circuit comprises two delay circuits R2 of equivalent delays making it possible to transmit to the multiplier M1, at the same time

  than a sample Xi, the previous sample Xi-1 and the next sample Xi + .l.

  The sample Xi is transmitted to the multiplier M1 on the one hand as it is and on the other hand multiplied by -1. As regards the samples Xi-1 and Xi + 1, only the signs of these samples are transmitted by the circuits sgnl and sgn2. The circuit M1 also receives a signal representing the value 0 of a sample. The circuit M1 is designed to compare the signs of the samples Xi-1 and Xi + 1 and to control, in application of the preceding table, the transmission to the output of the circuit M1, either of the signal + Xi, or

signal -Xi or signal 0.

  The processing carried out on the successive samples contained in the CCD of the TL6 detection device is integrated into the integrator I1. The information samples of the different tracks detected at a determined instant and contained in the CCD register of the detection device are therefore processed one after the other as has just been described and the results of the processing are integrated into the integrator II . Then, the samples of the different tracks detected at the following instants are treated in the same way and integrated following the previous ones. Integration can be

  do in the same integrator I1.

  The depth of the integrator depends on the average amplitude of the samples, their rate, and the desired response time. As an indication, for a rate of 15 million samples per second, a depth of the order of 20 bits makes it possible to obtain a bandwidth of the control loop of the order of 30 Hz in order to obtain a short latching time without affecting the accuracy of the system in normal operation, it is possible to change the time constant between the speed

  hooking and the tracking system.

  The error signal coded on 8 bits consists of the most significant bits of this integrator. It is then converted to analog quantity, filtered

  then adapted to control the electromagnetic motor.

  The result of the integration makes it possible to control the adjustment of the position of the detection device relative to the tracks of the strip, or what amounts to the same, in the case of an optical reading of the tracks, of adjusting the deflection of the beam. reading to the detection device. The detectors of the detection device being almost set on the centers

  tracks, it is then necessary to identify the numbers of the tracks read.

  We will now describe the relative position detection circuit DTR makes it possible to adjust the position of the detectors relative to the centers of the tracks read. However, as we have also seen (Figure 2), the detection device includes a larger number of detectors than there are tracks. All the detectors are nevertheless taken into account in the reading. It is therefore necessary to detect the detectors PHS1 and PHS2 not used in the reading of the tracks (FIG. 6). The DTA absolute position detection circuit makes it possible to detect these detectors and to identify the detector corresponding to the first track, that corresponding to the second track, etc. According to the invention, provision is made, from place to place on the recording medium, areas called preambles, containing signals

  special features to detect runways.

  In this preamble zone, at least some tracks, or even all tracks, contain identifying information allowing

indicate the number of the tracks.

  In these conditions, after having positioned the detection device in relation to the tracks, the detectors which read information

  of identification make it possible to identify the numbers of the tracks which they read.

  The system therefore identifies, on the one hand, the detectors useful for reading tracks and the numbers of the tracks they read. A preferred method according to the invention is to impose, in the preamble zone, signals having a continuous component

  positive or negative depending on the leads.

  In addition, these signals will preferably have characteristics required by other considerations. It is desirable, for example, for these signals to be poorly correlated from track to track so as not to fault the operation of the relative position detection circuit DTR. In addition, they must be able to be transmitted and exploited correctly by the system (spectral characteristics adapted to the channel, probability of transitions allowing a good operation of the phase locked loop). One method consists, for example, in using pseudorandom signals coded according to the 8/10 modulation code or a set of words specific to this code, with alteration of the coding, for this zone. In this system only the preamble area is affected by this coding. A variant may consist in maintaining this code in the areas of

  data, if code 8/10 is used in the data area.

  The code 8/10 with zero continuous component has a certain

  number of code words with non-zero continuous component (DC = + 2).

  The normal use of this code provides that a word with positive continuous component must be followed by a word with negative continuous component. If we violate this rule systematically or periodically, we generate a non-zero continuous component on the corresponding track. The integral of this continuous component is generally called DSV (Digital Sum Value). The idea is to make the DSV of each track moderately increasing or decreasing according to a predefined scheme for all the tracks and thus create an absolute track reference. When reading, we can distinguish tracks with increasing DSV from those with decreasing DSV. When writing, we make the DSV of the tracks alternately increasing on n tracks then

  decreasing on n following tracks.

  This involves a modification qualified as "alteration" on the algorithm of the initial coding. We will also speak of "modulation of the DSV"

  to qualify an increasing or decreasing variation of it.

  The method to make an increasing DSV consists in not compensating a coded word DC = +2 by a coded word DC = -2 and to alter the DSV by thus authorizing two consecutive coded words DC = +2. To make the DSV decreasing, on the other hand, it suffices to compensate twice a code word DC = +2 with a code word DC = -2 and therefore authorize two words coded with DC = -2 consecutive. It is also possible to systematically violate the encoding rule, which generates a stronger continuous component at the cost of a greater decoding penalty, if a continuous component is not

not reconstructed.

  To evaluate the continuous component on a track, we integrate the series of bits decoded on this track for a given time, and we

  examines whether the result is positive or negative.

  Rather than integrating the decoded bits track by track, it is possible to correlate the expected result with that which is

actually observed.

  For example, for an alteration of DSV of period 8 tracks, that is to say 4 consecutive tracks of positive DSV and 4 consecutive tracks of negative DSV, the algorithm is the following: S1 = sum [(bo + bl + b2 + b3) (b4 + b5 + b6 + b7)] S2 = sum [(b2 + b3 + b4 + b5) - (b6 + b7 + bo + bl)]

  o bi represents the signal samples of track i modulo 8.

  The value taken by S1 and S2 according to the position of the tracks is

shown in figure 7.

  It is then possible to deduce from the values of S1 and S2, by a correspondence table, the difference in position of the tracks with respect to their theoretical position. Such a table is as follows:

S1 -2 -1 0 1 2

S2

-2 _ 2 _

-1 3

0 4 0

1 5 7 _

2 6

  Conversely, having calculated the values of Sl and S2 we deduce the

offset tracks.

  To reduce the probability of error, we can normalize the

  values taken by S1 and S2, noting that IS11 +1 S21 = constant.

  It is therefore possible, by modulating the DSV of 8-track period, to obtain information on the position of the modulo 8-track tracks. If we want a larger range without uncertainty, we can increase this

period.

  A circuit performing this function is shown in Figure 8.

  Case of the use of any code words In the following, it is assumed that the code words are chosen randomly, but that a systematic violation of

  the encoding rule according to the tracks.

  The code has 256 words, of which 103 are non-zero DSV (DS = +2 or -2). For an equiprobable message, the probability of occurrence of a word with non-zero DSV is therefore 103/256. On average the DSV of this sequence is:

+ 2 * 103/256 = 0.805

  while the number of corresponding channel symbols is 10.

  The DC component generated by the code violation is then 0.08. This continuous component hardly penalizes the functioning of the system, and in particular the decoding (0.72 dB). The noise on the error signal can be evaluated as follows: For a measurement frequency of B = 100 Hz, number of symbols processed per integration period 1 / B: n = D / B o D is the gross bit rate The fluctuation in the content of the integrator is Vn. Typically, on a 64-track application D = 10 Mb / s and Vn = 316. In the case of a deviation from one track, the error signal is then, over the same period:

0.08 * D / B = 8000

Or an S / B of 29 dB.

  The optimal depth of the accumulator remains to be assessed. She

  depends on the desired detection time.

  Use of specific code words In a preamble, it is possible to provide specific code words. It is possible, in particular to use only words with non-zero DSV, which makes it possible to gain a factor 256/103 = 2.5, ie an improvement of 8 dB on S / B, with, however, a significant increase of the decoding penalty (2 dB), which presents no

  disadvantage in this preamble area.

  Use of words not belonging to the code table It is quite possible to choose words not belonging to the code 8/10, and their combination inside the track and between tracks, in order

  optimize the overall functioning of the system.

  Other methods of detecting the position of the tracks It is also possible to use other means of locating the tracks: for example, a specific signal (pure frequency or single sequence) on one or more tracks. The criteria for choosing between solutions are robustness (generally better ensured if the detection is done on a

  set of tracks) and the simplicity of the realization.

  Exploitation of the absolute position signal The detection system delivers a signal (flag) of position of at least one track, for example the zero track, in synchronism with the decoded bits of this track (figure 2). This information is used by the data rearrangement system (electronic formatting) to retain only the information corresponding to the tracks actually

  written (80/96 in our example).

Claims (7)

  1. System for reading a magnetic medium having several tracks of information readable in parallel, and comprising a detection device comprising at least as many detectors as there are tracks, making it possible to read a sample simultaneously and at regular intervals information on each track, said detection device comprising a parallel / serial shift register receiving in parallel the information samples read by the detectors at each reading instant and retransmitting them in serial form, characterized in that it comprises : * a processing circuit (M1l) receiving each sample of information (xi) to be processed from each track as well as the sample (x (i-1)) from a first neighboring track and the sample (X (i +1)) of a second neighboring track, and calculating the crosstalk to which the sample of information to be processed is assigned due to the neighboring tracks: * an integration circuit (11) receiving the value d e crosstalk thus calculated, integrating said values obtained at each reading instant, then integrating the values obtained at the following reading instants; + a relative tracking control circuit (CR) receiving the integration result of the integrating circuit (11) and supplying a tracking control signal from the detection device. 2. System according to claim 1, characterized in that said processing circuit comprises means making it possible to multiply the value of the sample to be processed: - by +1 when the sample of the first neighboring track is negative and the sample of the second neighboring track is positive; - par-1 when the sample of the first neighboring track is positive and the sample of the second neighboring track is negative; - by 0 when the samples of the neighboring tracks have the same sign; * an integration circuit (11) receiving the crosstalk value thus calculated, integrating said values obtained at each reading instant, then integrating the values obtained at the following reading instants; * a relative tracking control circuit (CR) receiving the integration result of the integrator circuit (11) and providing a tracking control signal from the detection device. 3. System according to claim 1, characterized in that the information carrier is read using a light beam which is transmitted to the detection device after reading the information carrier, and in that the circuit for relative tracking control (CR) allows to control a deflection device of the light beam according to the position of the detection device.   4. System according to one of claims 1 or 2, characterized in   that the detection device comprises a greater number of detectors than there are tracks to be read and in that it comprises: - an absolute position detection circuit (CTA) making it possible to identify the track read by each detector of the detection device; - a central control circuit (CC) controlling the operation of said processing circuit (M1) of said integration circuit (11) and of said tracking control circuit   relative (CR), then the absolute position detection circuit.   5. System according to claim 4, characterized in that it comprises means for identifying, in the information read by each   detector, one or more runway identity information.   6. System according to claim 5, characterized in that the tracks of the information medium includes preamble zones containing   said identifying information.   7. System according to claim 6, characterized in that the   preamble areas of the different tracks can be read simultaneously.   8. System according to claim 7, characterized in that the preamble zones have positive or negative components according to the tracks and in that a circuit makes it possible to detect the tracks with positive continuous components and those with negative continuous components. 9. System according to claim 8, characterized in that the tracks of the recording medium are distributed in alternating groups of positive and negative components. 10. System according to claim 9, characterized in that it comprises groups of four tracks of positive components which alternate with groups of four tracks of negative components and in that it comprises: - a first summing circuit (S1 ) adding the signs of the samples detected by a first group of four detectors (b0 to b3) and the inverse of the signs detected by a second group of four detectors (b4 to b7); - a second addition circuit (S2) adding the signs of the samples detected by the first two detectors of the first group of detectors and the last two detectors of the second group and the inverse of the signs of the samples detected by the other detectors of these groups detectors; - a table indicating the numbers of the tracks detected by said detectors as a function of the results of the additions   performed by the addition circuits.   1il. Recording medium comprising several tracks recordable in parallel, characterized in that said tracks each comprise a preamble area recorded or recordable in parallel, said areas containing information making it possible to locate the tracks moons compared to others.   12. Recording medium according to claim 11, characterized in that the preamble zones contain information with non-zero continuous components, the tracks being divided into groups of tracks containing information with positive continuous components which   alternate with groups of tracks with negative continuous components.