GB2060941A - Method for displacement of a movable system with respect to a data carrier and a device for carrying it out - Google Patents

Abstract

Method for the displacement of a movable system with respect to a data carrier recorded on a plurality of tracks whose addresses are written on the carrier within a plurality of reference zones, each track being associated with at least one zone, the system comprising a head for reading the data. At determined sampling instants a set or instructed acceleration gamma c is calculated as a function of the address ADLj read by the head at these same instants, and this is compared with the measured acceleration gamma of the system and the motor is controlled as a function of the result of the comparison between gamma c and gamma . Applicable to the displacement of heads for reading disc memories. <IMAGE>

Description

SPECIFICATION Method for displacement of a movable system with respect to a data carrier and a device for carrying it out The present invention relates to a method for the displacement of a movable system with respect to a data carrier and a device for carrying out the method. It is more particularly applicabie to the displacement of read/write heads of disc memories used in data processing systems.

In such systems, magnetic disc memories are used more and more frequently because of their storage capacity and of the relatively short time taken by the magnetic writestread heads to access an item of data contained at any point on the disc from the moment they have received the order to access this information.

It is known that magnetic discs carrying the data in coded form on concentric circular recording tracks whose width does not exceed a few hundreths of a millimeter and which are situated on both their faces.

The codes most frequently used are binary codes.

The tracks are identified giving them a serial number j, j being a whole number varying from 0 to N - 1, N being the total number of recorded tracks.

The coded expression of the serial number j of the track is called the address.

To enable the reading or writing of data, the magnetic heads are arranged on each side of the discs at a distance of a few tenths of microns from the latter.

The magnetic discs are driven by an electric motor at a constant speed of rotation.

In current practice and more particularly in the case of memories which only comprise a limited number of discs (generally less than 4 or 5), the data is recorded on each of the disc faces in the following manner. A maximum space is reserved for the recording of information or data intended to be processed by the data processing system to which these memories belong, this data being called "data to be processed" for simplification. A minimum of space is reserved, on the one hand, for recording the addresses of the tracks and on the other hand for recording data necessary to control the position above the tracks of the magnetic head or heads associated with this face. There will be designated in the following under the name of track identifying data both the addresses of the latter and the data for servo-control of the position.

For simplification only one face of a disc will be considered and it will be taken that only one magnetic head is associated with it. The latter reads (and/or writes) both the data to be processed and the addresses of the tracks and the servo-control position.

In current practice, as has been described for example in patent application No. 4116/77 filed on 1 st February 1977 in the name of Compagnie Honeywell Bull under the title "Method of writing addresses on a magnetic record carrier", the data contained on each face of the disc are preferably distributed over equal and adjacent circular sectors So, S1.. .., S,.. ..., S,, 1. Generally a face of the disc is divided into several tens of sectors (most often 40 to 50).

When the face of the magnetic disc passes in front of the magnetic head which is associated with it, the sector S0 is read by the head before the sector S1, the sector S1 before the sector S2 and so on. It is therefore said that sector S0 precedes sector Si, that sector S1 precedes sector S2, that sector S, precedes sector Sl+l, etc..

More generally, when two items of data Ik 1 and 1k which follow one another on the same track of serial number j of the said face are considered, it is said that the item of data 1k-i precedes the item of data 1k if it is read by the head before the latter, or again that the item of data 1k follows the data Ik~ 1 The reasoning is also valid for several groups of data Gk and Gk-1.

Each sector S, is divided into two unequal areas. The larger area comprises the data to be processed whilst the smaller area comprises information for identifying the tracks. For each sector, the smaller area is divided into several zones called reference zones. equal in number to that of the tracks, each track being associated with one and the same zone.

It will be recalled that the English word "bit" designates both a binary Figure 1 or 0 or any representation of this figure either in the form of magnetic recording, or in the form of an analog or logic electric signal, a logic signal being able to assume only two values called "logic 0" or "logic 1" and an analog signal being defined as a signal whose voltage may vary continuously between two positive or and negative limiting values. For simplification, any item of data contained on the disc will be designated in the following under the name of "bit" In particular, the items of data for identifying the tracks will also be called track reference bits.

To shorten the time taken by the head to access any "item of data to be processed", it Is necessary principally that the head can move from one track to another in the shortest time possible and be positioned precisely with respect to this latter.

Thus, devices are known which permit displacement and positioning of the head in response to these requirements. Certain of them called "bang-bang" use an electro-dynamic motor of the "VOICE COIL" type comprising a coil displacing linearly within a permanent magnet of cylindrical shape. This coil is mechanically connected to a carriage which carries the magnetic head, by means of a suspension arm.

A movement comprising two phases, one of acceleration and the other of deceleration is Impressed on such a device for displacement and positioning of the head. During the first phase a constant current (for example positive) is applied to the coil of the motor. Under these conditions the speed law of the carriage (and therefore of the heads) can be assimilated to an increasing linear function of the displacement time of the latter.

The curve representing the speed as a function of the momentary position of the carriage is an ascending arc of a parabola, the speed increasing as a function of the position.

During the second phase of the movement, which is a deceleration phase, an inverse current (for example negative) is applied to the motor. The speed of the carriage being therefore a decreasing linear function of the time, the curve representing the speed as a function of the position occupied by the carriage is an arc of a parabola, the speed decreasing as a function of the position. At the end of the second phase, the speed of the carriage and the space which remains for it to travel through should be sufficiently small for the heads to be stopped above the track selected.

Patent application No. 531 99!75 filed on the 20th December 1976 by the Compagnie Honeywell Bull under the title "Method for displacement of a movable system with respect to a data record carrier and a device for "carrying it out" describes and claims a simple and advantageous embodiment of a device for displacement and positioning of a movable system whose principles of operation are set out above. It also describes and claims the method employed by this device.

In this method, where the address of the track is the only data controlling the current in the coil of the electro-dynamic motor which drives the movable system, the magnetic read/write head is displaced from a departure track A to an arrival track B, of which the addresses are furnished by a circuit for controlling the addresses belonging to the memory which contains the disc associated with the said head. During the movement acceleration phase, the motor is supplied by a constant current from the track A to a track C where the current is reversed, the deceleration phase taking place from the track C.

The characteristics of this method are the following: - the addresses of addresses of the tracks are recorded on the disc in reflected binary code; - the address of the track C is calculated as a function of the addresses of the tracks A and B, these 3 addresses being expressed in weighted binary code; - the addresses of the tracks read by the magnetic head during its displacement are memorised and then transcoded into weighted binary code; - during the acceleration, these transcoded addresses are compared with the address of the track C; - the deceleration takes place up to the moment where the speed of the movable system is less than a minimum threshold V0 calculated from the read and transcoded addresses; - the address of the track opposite which the magnetic head is immobilised is compared to the address of the track B; - a new displacement takes place if these addresses are different.

It appears clearly from the preceding that the two phases of movement takes place in an open loop or again in free operation, that is to say without servo-control. The following results: - a) under the action of various parameters such as direction of displacement, temperature, characteristics of the motor, (inductance and resistance of the coil), coefficient of force, etc , and under the action of external disturbances, for example, dry and viscose friction, effects of weight due to the more or less large inclination of the disc memory, external vibrations, the distance which remains to be travelled by the head from the moment when the speed has dropped below the threshold V0 is very variable (even for the same starting track A and the same arrival track B), which necessitates several successive repetitions in order to attain the track B from which follows an increase of the time taken by the head to access the "data to be processed" recorded on this same track, from the moment when they have received the order to access it (for simplification this time is called "access time"); - b) for small distances, that is to say when the distances apart, expressed as the number of tracks, between the starting A and arrival B tracks are comprised between 1 and 5, the method described above needs to be modified; - c) it is difficult to obtain small access times.

The present invention enables these disadvantages to be remedied by servo-controlling in acceleration the movement of the movable system comprising the magnetic head, by the calculation at determined sampling instants of an acceleration instruction *,c function of the address the track opposite which is found the head at these same instants, and by the comparison of','c with the measured value , for the acceleration of the said system. The voltage applied to the terminals of the coil is a function of the result of this comparison.

With respect to the method of displacement which Is described in the above-cited patent application, the method according to the invention enabies, on the one hand, a large reduction in the time taken by the write read head to access any item of data recorded on the face the disc which Is associated with It and, on the other hand, that the read head comes opposite the arrival track B in a single movement (without there being any need for several successive movements to reach it, as in the method according to the prior art) According to the invention, the method for displacement of a movable system with respect to a carrier for data recorded on a plurality of tracks whose addresses are written on the carrier within a plurality of reference zones at least equal in number to that of the tracks, each track being associated with at least one zone, the system being actuated by an electric motor and comprising at least one data read head which Is displaced from a starting track A to an arrival track B of addresses AD1, the addresses of the tracks read by the head during displacement being designated by ADL1, is characterised in that: 1) - at determined sampling instants one acceleration Instruction;e Is calculated as a function of the address ADL read at the same instants.

2)the acceleration ? of the system is measured.

3) - the accelerations c and are compared.

4) - the motor is controlled as a function of the result of the comparison between yc and (.

In the preferred embodiment of the invention, the method is characterised in that at the said sampling instants: a) the separator E1 = AD - ADLI is calculated b) a non-linear function f(E1) of zi is determined. c) there is calculated the speed v of the movable system as a function of the difference between the addresses read by the head at the instants of sampling separated by determined time intervals. d) the acceleration instruction ye is calculated proportionally to the sum (flue1) - v).

Other characteristics and advantages of the present invention will appear in the following text which is given by the way of non-limiting example and with reference to the acccompanying drawings; In the drawings, Figure lisa set of Figures lato 1ewhich show a preferred example of distribution of data on one face of a magnetic record carrier such as a magnetic disc: Figure2 is a set of Figures 2a, 2b, 2c which illustrates a preferred method of writing the addresses of tracks of one face of a magnetic disc within a reference zone of this same face: Figure 3 is a block diagram of the device for the displacement of a movable system with respect to a data carrier such as a magnetic disc, carrying out the method according to the invention: Figure 4 is a curve showing the variation the function f(E,) as a function of the address separation E1.

Figures is a curve showing the variation of the speed of the movable system as a function of the time; Figure 6shows the variation, as a function of the time, of the supply voltage of the coil of the linear electro-dynamic motor which drives the movable system; Figure 7 is a more detailed block diagram of the displacement device employing the method according to the invention; Figure 8 illustrates the accuracy with which the address ADLj of a trak of serial numbers j is determined; Figure 9 shows how the average measured speed is estimated with an estimation delay 0 with respect to the actual speed of the movable system.

In order better to understand the principles of constitution and operation of the device for displacement of a movable system with respect to a record carrier employing the method according to the invention, it is useful to recall several points illustrated in Figures 1a to 1e and 2a to 2c showing, on the one hand, how the data are distributed on the surface of a magnetic record carrier which is preferably a magnetic disc (Figures la to le) and, on the other hand, a preferred method of writing data within a reference zone of this magnetic disc (Figures 2a, 2b, 2cl.

In Figure 1 a, is considered one face of a magnetic disc D turning in the direction of the arrow F, whose useful recording surface is delimited by the circles d1 and d2. It is assumed that it is associated with a single magnetic write read head TEL. There are defined on this disc nO equal and adjacent circular sectors So, S1, S, ..SnO1. As can be better seen in Figure 1 b, each sector Si is divided into two parts SAD, and SDO, wherein are recorded respectively the addresses of the tracks and the "data to be processed" by the data processing system with which the disc memory containing the magnetic disc D is associated.

The surface of the part SAD, is very much less than the surface of the part SDO,.

Figures 1 c and id show in more detail the way in which the parts SAD, of the sectors S, are formed. They are an enlarged view of the part SAD; of the sector S, comprises within the circle C.

Each part SAD, of a sector S, is divided in N zones ZIP,0 ZRP,+ ZRP,N 1, (N being the number of magnetic trucks of the magnetic disc D).

In Figures ic and id. only the first five zones ZRP@@ to ZRP,4 have been shown for simplification.

The boundaries between the different zones ZRPij are the circular axes Ax, of the magnetic recording tracks. To each magnetic track, of serial number j of axis Ax is associated the zone ZRP,, Thus the reference zone ZRP@@ is associated with the track of serial number 0, reference zone ZRP,I is associated with the track of serial number 1, and soon.

It is recalled that the magnetic read and or write heads comprise a magnetic circuit round which is disposed a wiring and which comprises an air-gap. So that the "data to be processed" of a track of serial numbers j of magnetic axis Ax are read by a magnetic read head TEL with the maximum accuracy, the latter remaining motionless with respect to the same track during the time necessary for the reading of this data, It is necessary that Its air-gap is perfectly centred on the magnetic axis Ax, which is the boundary between the two reference zones ZRP and ZRP , -11. It could therefore also be said that the magnetic read write head TEL is disposed astride the two zones.

To simplify Figure id, the reference zones ZRPi have been represented by rectangles. Each of these zones contains the address of the track to which it is associated. As can be seen In Figure id, the zone ZRP,o contains the address of the track of serial number 0, the zone ZRP@@ the address of the track of serial number 1, the zone ZRP, the address of the track of serial No. 2 and so on.

The address of the tracks Is written according to a reflected binary code. called the GRAY code. The description of a code of this type Is given for example in the book by H. SOUBIES-CAMY published by an Editions DUNOD In 1961 pages 253 to 254. An example of the writing in GRAY code of two successive addresses, those of tracks 124 and 125 is given in Figure le.

This example illustrates the main characteristic of the GRAY code, i.e. that two successive addresses are distinguished by the change of only one bit between them. Thus, the two addresses 124 and 125 written in GRAY code differ by the last bit, equal to 0 for the track 124 and equal to 1 for the track 125.

In Figure 2a is considered a reference zone ZRPij of a sector Si, the direction of movement of the disc D being indicated by the arrow F. As has been described in patent application No. 7829847 filed on 19th October 1978 under the title "Method of writing data on a magnetic recording carrier", by the applicant company, the address of the track is contained in a part PAD of the latter, the rest of the zone containing principally data for servo-control of the position of the head TEL on the axis Axj of the tracks of serial No.

The reference zone ZRPj, is preceded by a zone ZB-,1, called the "white zone" which separates it from the part SDOi of the sector S containing the data to be processed.

The magnetic induction is uniform in the zone ZB,i and is, for example, negative as indicated in Figure 2a.

It is known that, to record a data on a magnetic disc, a succession of small magnetic barriers (these dimensions of which are of the order of a few microns) called elementary areas are created on each track of the disc, these areas being of variable length and distributed over the entire length of the track and having alternatively magnetic inductions of the same module and of opposite sense, of direction parallel to the surface of the disc.

The start of the reference ZRP,j is indicated by the reference DZ,i. It is constituted by a change in sense of the magnetic induction between the zone ZB;i where the induction is negative and the first magnetic area DMj of the zone ZRP,j where the magnetic induction is positive.

In the rest of the text a change in direction of the magnetic induction will also be called magnetic transition.

A magnetic transition can be of two different types namely: - when the face of the disc travels in front of the magnetic head T and the latter sees going past successively an elementary magnetic area of negative magnetic induction and then an elementary area of positive magnetic induction, the corresponding magnetic transition is called positive.

- when, on the contrary, the magnetic head T sees going past successively an elementary area of positive induction and then an elementary area of negative induction, the magnetic transition is said to be negative.

The part PAD which comprises the addresses is composed of m identical elementary cells (12 in the embodiment shown in Figure 2a) of length L, i.e. the cells C0, C1 , Ck , C11, each cell containing one bit of the address. Any bit Bk of the address contained in a cell is defined by the presence or absence of a double magnetic transition, the first magnetic transition Tik being of opposite sign to the second position T2k. For example, the first transition Tik is positive (see Figure 2b) whilst the second T2k is negative. The coding of the address bits ADE, of the track of serial number j contained in a reference zone ZRPj, is selected for example such that the bit Bk is equal to 1 in the case of the presence of the double magnetic transition, whilst it is equal to 0 in the case of the absence of the latter, this absence being translated by a uniform magnetic induction, for example negative, in the cell containing this bit of zero value (see Figure 2b). For simplification, in the following the absence or presence of a double magnetic transition will be designated under the English name of "dibit" Figure 2c shows the analog signal delivered by the magnetic head TEL when a cell Ck passes in front of it.

When the bit Bk is equal to 1 the signal delivered by the head TEL is composed of two analog pulses of opposite sign whose amplitudes are equal, in absolute value, to AMP. When the bit Bk is equal to 0, the voltage of the signal supplied by the head TEL remains nil. As can be seen in Figure 3 which represents the device for carrying out the method for displacement of a movable system with respect to a data carrier according to the invention, the movable system SYSMOB to be displaced is constituted by the magnetic readswrite head TEL which Is associated with a carriage CHAR to which it is mechanically attached.

The aim of the device for carrying out the method according to the invention is to displace, in a single course in the shortest time possible, the magnetic write read head TEL from a starting track A to an arrival track B of address AD,. The movement of the head TEL Is controlled by a non-linear second order differentially equation of the type: f(#1)+d#1+1d#1=0(1),#1 and f(#1) dt C2dt2 being respectively the non-linear variable and function defined above, fl,) being increasing, C2 being a constant.

Here 12 = dt-, dt = -v, v being the speed of the head TEL, and #3 = d#1 dt2 = - Is where, Is the acceleration of the head TEL.

The equation (1) could therefore be written: f(#1) + #2 + #3,C2 = 0 (1') In a preferred embodiment of the invention, the method for displacing the movable system SYSMOB with respect to the face of the disc D comprises the following operations: 1) - at determined sampling instants. regularity spaced in time, the time Interval separating these sampling instants being equal to T seconds. the acceleration Instruction, is calculated In the following manner: a) the address ADL Is determined and the separation t, IS calculated. b) the corresponding function f(#1) Is determined, a known function perfectly deterrnined in advance; It could also be said that fli ,) IS a function of the address ADL, c) the speed v of the system SYSMOB is calculated as a function of the difference between the addresses ADL (nT + koT) and ADL (nT) which are the addresses ADLj read at the sampling instants: tn = nT and tk0 = nT + K0T, n and k being whole numbers. d) there is calculated the acceleration instruction yc such that yc/C2 = (fIr1) - v); it is seen that yc is a function of ADLj.

2) - there is measured the acceleration &gamma; of the system SYSMOB which is divided by C2.

3) - the difference ("c - y)/C2 = #(&gamma;C2) is calculated.

4) - the coil of the electro-dynamic motor ML is fed by a voltage of which the sign depends on the sign of the difference A(yC2).

The different essential constituent elements of the device for carrying out the method according to the invention are; - the electro-dynamic motor ML - the circuit CIRCAD for determination of the read address ADL - the address control circuit GESTAD - the means ACCEL for calculating the acceleration instruction &gamma;c - the means MES for measuring the acceleration y - the comparator COMP effecting the comparison between the set acceleration yc and the measured acceleration y - the voltage supply generator ALIM for the coil of the electro-dynamic motor ML The circuit CIRCAD: a) receives the analog signal ST supplied by the magnetic write/read head TEL when the dibits of data contained in the part PAD of a zone ZRP,i move past in front of it, the signal ST being composed of a set of analog pulses. b) it transforms the latter into a set of logic pulses which form the address ADGj, expressed in GRAY code, of the track of serial number j associated with the reference zone ZRPij. c) it then transcodes the address ADGj into an addressADL1 expressed in weighted binary code, such codes being described in the book by SOUBIES-CAMY previously cited. d) its supplies to the means ACCEL for calculation of the acceleration instructed or set, on parallel tracks, the address ADL, with a sampling frequency of F = 1/T, the sampling period T being equal to the time separating the passage of two parts PAD of two reference zones ZRPii and ZRP(i+1)j associated with the same track of serial number, the first of which precedes the second. In other words, it can be said that the addresses ADLi are supplied by the circit CIRCAD every T seconds.

The set acceleration means ACCEL comprise (Figures 3 and 7): - the subtractor SOUS calculating the quantity E1 = ADFf - ADL - the generator GF of the function f(t 1) - the calculator CALVIT for determination of the measured speed vrn - the adder ADDIT (see Figure 7) - the digital-analog converter CDA (see Figure 7) - the device COMPRET for compensation of the average delay H of estimation of the measured speed vrn with respect to the actual speed v of the head TEL - the adder ADD which supplies the set acceleration &gamma;c (ADD, ADDIT, CDA are represented in the form of a single block, for simplification, in Figure 3).

The subtractor SOUS receives on the one hand, the addresses ADL from the circuit CIRCAD and, on the other hand, the address AD, from the track B, sent by the address control circuit GESTAD of the data processing system of which part is the disc memory containing the disc D. The address AD is expressed in the same weighted binary code as the address ADL1.

It is clear that the subtractor SOUS, receiving a new address the track of serial number j at approximately half a track.

The quantity 1 q therefore represents the distance travelled by the head TEL during a time interval equal to (k0 x T) seconds.

The calculator CALVIT determines the measured speed vrn according to the formula vm = 1 q/k0T; the speed vrn with a changed sign is transmitted in binary form to the adder ADDIT.

For reasons which will be explained in more detail later, it is shown that the measured speed calculated at the instant (nT + K0T) is not equal to the actual speed v of the magnetic head TEL at this moment, but is equal to the speed of this same head at the instant (nT + K0T) - # with 40 equal to (K0 + 1) T/2, # being called the average estimation delay.

The device COMPRET for compensation of the average delay # has for its purpose to compensate for the effects of the latter on the measurement of the speed Vm; it receives the signal y and supplies a compensation signal &gamma;F.

If the quantity (vm + YF) = v is called the estimated speed and #v the speed difference v - v = v - vrn - YF, the characteristics of the delay compensation device COMPRET are established so that the speed difference #v is minimum, that is to say nil. Thus it could be said that the estimated speed v is practically equal to the actual speed v of the magnetic readlwrite head TEL. It is shown that this result is obtained for a value of YF = y x G where G is the transfer function of the device COMPRET which is preferably a filter. The signal &gamma;F@ with a changed sign, is sent in analog form to the adder ADD.

The adder ADDIT calculates the sum S = (-vm + f(#1)) expressed in binary form, which is sent to the digital/analog converter which sends it in the form in an analog signal to the adder ADD.

This latter therefore receives analog signals (-vm + f(#1)) and -YF. It delivers the acceleration instructed &gamma;c at practically a constant. In effect: -vm + f(#1) - &gamma;F = -(vm + &gamma;F) + f(#1) = f(#1) - v = f(#1) + #2 = -#3/C2 = &gamma;c/C2 The adder ADD sends the signal y c/C2 = - EiC2 to the comparator COMP.

The means for measuring the acceleration MES deliver a signal y according to the following principle; it is shown that the acceleration of the movable system driven by the linear electrodynamic motor ML is proportional to the current circulating in the coil. It is therefore sufficient to measure this current i and to multiply it by a proportionality coefficient to obtain the signal y, sent on the one hand to the compensation device COMPRET and, on the other hand, multiplied by ii'C2 (by the multiplier MUL) sent to the comparator COMP.

At the output of the comparator COMP is collected a signal lyc - y)/c2 = (E3 - E3)/C2 = #(#3/C2) which controls the supply generator ALIM.

If #(#3/C2) is positive. the supply generatorALlM supplies a voltage + Uo to the coil of the linear electro-dynamic motor ML.

If #(#3/C2) is negative, the supply generator ALIM supplies a voltage - Uo to the coil of the linear electro-dynamic motor ML.

In order to clarify the principles of operation of the device for carrying out the method according to the invention, the comparator COMP and the adder ADD are considered as two separate operational elements but it is clear that, in practice, they can constitute one and the same element accomplishing successively the addition of the signals (-vm - &gamma;F) and f(#1) to obtain &gamma;c/C2, then the comparison between the signals &gamma;/C2 and .,'c C2 (between r 3.C2 and #3,C2).

Figure 5 shows how the speed of the movable system SYSMOB evolves during its displacement between the tracks A and B.

From track A to track C the coil of the motor ML Is permanently supplied with a positive voltage signal + +Uo (see also Figure 6) so that the motor ML is saturated.

Figure 5 shows that, between the points A and B (corresponding to the tracks A and B), that is to say between instants tA and te. the curve X of speed variation is more or less of expoential form, the speed remaining less than a speed Vm In effect, for sufficiently large values of #1 1 it Is estimated that. around each abscissa point fl, (see also Figure 4) one has; f(t1) = a &alpha;#1 (2) with ii = df(#1) d#1,&alpha; being very small. It is then seen that the movement of the movable system is controlled by a differentially equation of the form; #2 + 1 C2 d#2.dt = consatnt (3) the solution of which is of the type; r2 = B1 (1 - e c2') (4) In other words it can be said that. between the tracks, A and C, the movement of the movable system SYSMOB is speed-regulated.

When the head TEL approaches the track B (smaller deviations of the address t 1), the approximation given by the equation (2) Is no longer valid and the movement of the movable system SYSMOB undergoes a regulation defined by the non-linear second order differential equation (1) already cited.

The speed curve of the movable system SYSMOB is therefore the curve )2, starting from point C (instant t); it can then be said that the system SYSMOB is caused to slide on a trajectory corresponding to the non-linear second order differential equation (1).

Figure 5 also shows the curve of variation of the speed as a function of the time when the head TEL Is displaced between a track A' and track B, the distance between the latter being larger than the distance between tracks A and B. The speed variation Is then given by the curves # # (between points A and Cl and by the curve #2 between the points C' and B).

During the second part of the movement that is to say between C and B or C' and B (refer to curves '2 or #2), the supply voltage U of the coil of the motor ML which is given in Figure 6 by the curve 7x'3 in broken lines, is constituted by a set of pulses of negative and positive voltage of variable duration. Its average value is given with a curve #3 in solid line.

If it is assumed that the voltage of the U0 is positive. the average value of the voltage U between the instants tc(or t'c) and tB is negative.

The address determination circuit CIRCAD comprises as can be seen in Figure 7; - the threshold circuit GS; - the register transcoder TRANSCOD; - the sample generator ECHANT supplying sampling pulses every T seconds; that is to say defining the sampling instants.

The thresholds circuit GS receives the signal ST and transforms the set of analog pulses constituting this latter into a set of logic pulses by means of two thresholds S1 and S2. If it is assumed that the absolute value of the average amplitude of the signals delivered by the head TEL corresponding to the bits equal to 1 (presence of double transition referred to in Figure 2), is equal to AMP, one has, in a preferred embodiment of the invention; S1 = 0,25 x AMP (5) and S2 = 0,75 x AMP (5') The method of determination of the value of the bits by the circuit GS is then the following; - considering two adjacent reference zones ZRP,, and ZRPi(j+1), the corresponding track addresses written in the parts PAD of these zones being respectively ADE and ADEj+1, and considering two cells of the same series K within these two zones. namely cells C,1 and Ckii1i. The bits corresponding to these two cells are respectively B,1 and Biii1. Due to the fact that the addresses ADE and ADEj+l are written in GRAY code, three cases can arise; - Case 1: the bits Bkj and B,11 +1) are nil. The voltage of the signal ST is nil and as a result less than the threshold S1. The circuit GS then supplies a signal equal to "logic zero" and this, whatever the position occupied by the read head TEL when it moves from the position POS1 where its air-gap is situated opposite to the zone ZRPij (see Figure 8 where the air-gap has been shown by a rectangle of which the length is very much greater than the width) to the position POS3 where this air-gap is situated above the zone ZRP,1111, passing by the position POS2 where this air-gap is situated astride these two zones, that is say, centred on the axis Ax of the track of serial number.

- Case 2: the two bits Bk, and Bk(j-1) are equal to 1. The voltage of the signal ST has a positive amplitude and a negative amplitude whose absolute value is equal to AMP, that is to say greater than S2. The circuit GS then supplies a signal equal to "logic one" whatever the position occupied by the read head TEL between the positions POS1 and POS3 (see also Figure 8).

- Case 3: it is assumed that Bkl is equal to 0 and Bk(j-1) is equal to 1. The two addresses ADE and ADEj+1, differing from each other by a single bit, this third case therefore only occurs for a single bit of the same type for two adjacent reference zones. Consider then the evolution of the absolute value of the amplitude of the signal ST (see Figure 8). The distance between the position POS1 and POS3 is equal to the width of a zone ZRP,i, itself equal to the width 1 p of a track. This distance 1 p is also called the pitch between the tracks. It is clear that, when the head TEL is displaced continuously between the position POS1 and the position POS3, the absolute value of the amplitude of the signal varies continuously from 0 to 100% of AMP. It is said in this third case that the signals ST Is an ambiguity and that it corresponds to an "ambiguity bit", the amplitude of the ambiguity varying as a function of the position x occupied by the head between the positions POS1 and POS3. Let A(x) be this amplitude. It Is seen that if x Is less than 1p/4, A(x) Is less than 0.25 AMP = S1.

If it is seen on the other hand that if x Is greater than 3 1 p/4, A(x) is greater than 0.75 AMP = S2.

Finally if Alx, is comprised between S1 and S2 that is to say between 0.25 AMP and 0.75 AMP, one has 1p,4 < x < 3 1 p 4.

The threshold circuits GS supplies an address read in GRAY code GRAY i.e. the address ADG, or ADGj+1, the register-transcoder TRANSCOD receiving from the threshold circuit GS according to a frequency 1 T equal to that of the sample pulses delivered by the generator GEN. the addresses ADGj read in Gray code and transcoding them into weighted binary code. The register TRANSCOD which is controlled by the sampling generator ECHANT therefore supplies every T seconds, on parallel tracks. the address ADL expressed in weighted binary code and transmitted to the subtractor SOUS and to the speed calculator CALVIT.

If it Is convenient to define after conversion of the address ADGj read in GRAY code into an address ADL, read in weighted binary code, a binary weight a 1 (j) such that; if { x < p/4 , x > 3 1p/4 A(x) < 0,25AMP ,A(x) > 0,75AMP AMP= 0 16) and 1p/4 < x < 3 1p/4 if 0,25 AMP < Alx)0,75 AMP a,(j) = 1 It Is known to represent any position whatsoever of the head TEL opposite the face of the disc which Is associated with It by an address quantified as a half step (half width of the track.) Thus If It is taken that the serial number I Is equal to 124 (it Is assumed therefore that the head Is displaced from 124 to track 125). if x < 1p/4 that is to say if a(x) is less than S1, it is considered that the head TEL occupies the position 124.

If x > 3 1 p/4, that is to say A(x) is greater than S2, the head occupies the position 125.

If 1 p/4 less than x which is itself less than 3 1 p/4 it is convenient to say that the head occupies the position 124.

Under these conditions, the position of the head TEL on the disc is expressed by the address ADLji such that, ADLj = a-1 (j).2-1 + ao(j) 2 + a1(j) 2 + ...... an(j) 2n with a1(j), a2(j).. ..an(j)#{0,1} (7) weight 1 = 1 ip2 If, as has been described, the final position occupied by the head TEL is such that the latter is positioned astride the magnetic axis Axf of the address track ADf with; ADf = 1.2-1 + ao(f) 2 + an(f)2n with aO(f), a1(f) anif) belonging to (0,11 (8) The separation F1 = ADf - ADLj, expressed as a half step, can be calculated in binary form in the following manner: #1 = #1-1(j).2-1 + #10(j).2 + #11(j)2 + .....#1n(j)2n with#1i(j)#{0,1} (9) The accuracy of determination of the head position and of the separation r1 is equal to 1 p/2 = q.

The speed calculator CALVIT comprises (see Figure 7); - the circulating memory MEMOCIRC, - the subtractor-divider SU BDIV, - the block device BLOC The circulating memory MEMOCIRC receives every T seconds the address ADL(nT + koT) and delivers the address ADL (nT) to the subtractor-divider SUBDIV. This latter also receives the address ADL (nT + koT.) The circulating memory retains all the values of the address read between instants (nT) and (nT + koT,) that is to say the addresses ADL (nT), ADL (nT + T), ADL (nT + 2T),. ADL(nT + koT). The subtractor divider SUBDIV calculates the speed vm by determining the difference ADL (nT + KoT) - ADL (nT) and by dividing it by the quantity KoT (operations carried out at each sampling instant, that is to say every T seconds.) The blocking device BLOC blocks the value of vm = ADL (nT + koT) - ADL (nT) / koT during a time interval equal to T seconds.

The determination of the average estimation delay H which is shown in Figure 9, is based on the following principle; the time interval which separates the instants nT and nT + koT, that is a time interval equal to koT is sufficiently small (a few milli-seconds) for the variation is actual speed v of the head TEL to be considered during this time interval, is linear as a function of the time. The corresponding variation curve 72is shown in Figure 9. The symbols to, t1, t2, t3, t4, tS, t6 etc.. . designate respectively the instants. . nT, nT + T, nT + 2T, nT + 3T, nT + 4T, nT + 5T, nT + 6T. etc.. .. and it is taken that ko is equal to 4.

At the instant t4 the subtractor-divider SUBDIV calculates the value vmi = (ADL(nT + 4T) - ADL (nT)) ,' 4T.

This value is blocked for T seconds by the blocked device BLOC that is between the instants t4 and t5. At the instant tS the value ADL (nT 4 5T) - ADL (nT + T)),4T = vm2 is calculated, a value which is blocked for T seconds between the instants tS and t6. Similarly at the instant t6 one calculates the value vm3 = (ADL(nT + 6T) - ADL (nT + 2T))/4T which is blocked for T seconds between the moments t5 and t6. The values vm1, vm2, vm3, therefore represent the speeds measured at the instants t4, ts, t6. It is clear that the process of determination of the measured speed vm is identical to that which has just been described, both for instants before t4 and for Instants after t6. The curve representing the variation of the measured speed vm as a function of the time is the curve @5, The variation of the average speed vm as a representative curve 6.

Due to the linear evolution of the actual speed v as a function of the time, the speed vm measured in the way indicated above at instants t1, t2, t3, t4, t5 etc.. . Is equal to the actual speed v measured at the instant (nT + ko T.2) (see Figure 9 and compare the curves 7'4 and @5). Thus the speed measured at the instant t4 is equal to the actual speed at the instant t with t2 = (t4 + to)/2 = to + (t4 - to)/2 = to + koT 2 = to + 2T 2T.

This results from the fact that when the speed evolves linearly as a function of the time, the average speed between two determined instants Is itself equal to the speed measured at the middle of the time intervals separating these said instants. As the value of the average speed van cs blocked for T seconds. it is clearly seen from examination of Figure 9 that the average estimated delay H is equal to koT 2 + T 2 = (ko + 1) T 2.

The optimum value of k,) is determined in the following manner; it is known that v= = 1 q koT and that the accuracy of determination of the quantity 1 q is equal to q.

The result is than an error called "quantification error" Fq exists in the determination of the measured speed vm equal to q/koT. To this quantification error should be added an error #@ due to the average estimation delay) = (k" + 1, T 2. The result is #@ = &gamma; t) is obtained lin effect = dv/dt, that is dv = ydt).

If a function Q called the "cost function" Is defined such that one has Q = @q + r,i @ It is seen that in deriving this function, there exists a value ko = 1 T x #2q &gamma;' which minimises the cost function Q. It is found that ko = 4, in the embodiment described here.

Claims (8)

1. Method for displacing a movable system with respect to a data carrier recorded on a plurality of tracks whose addresses are written on the carrier within a plurality of reference zones, at least equal in number to that of the tracks, each track being associated with at least one zone, the system being driven by an electric motor and comprising at least one data read head which is displaced from a starting track to an address arrival track ADf, the addresses of the tracks read by the head during displacement being designated by ADLi, characterised in that; 1) at determined sampling instants, a set or instructed acceleration Yc is calculated as a function of the address ADLj read at the same instants.

2) the acceleration fir Of the system is measured, 3) the accelerations Yc and glare compared, 4) the motor is controlled as a function of the result of the comparison between y, and y.

2. A method according to claim 1, characterised in that at the said sampling instants, a) the separation E1 = ADf - ADLj is calculated b) a non-linear function f(#1) of E1 is determined c) the speed v of the system is calculated as a function of the difference of the addresses read by the head at sampling instants separated by determined time intervals do the set acceleration 'Ic is calculated proportionally to the sum (f(#1) - v).

3. A device for carrying out the method according to claim 1 or 2, comprising a motor voltage and/or motor current supply generator, an address control circuit supplying the address ADf, a circuit for determination of the address read ADL characterised in that it comprises; - means ACCEL for calculating the set or instructed acceleration Yc receiving the addresses Ado and ADL supplied respectively by the address control circuit and the circuit for determination of the address read, means MES for measuring the acceleration &gamma; of the said movable system SYSMOB, - a comparator COMP effecting the comparison between the set or instructed acceleration Yc and the measured accleration y, - the comparator output signal controlling the said supply generator.

4. A device according to claim 3, characterised in that the means ACCEL for calculation of the set or instructed acceleration comprises; - a subtractor SOUS receiving the addresses ADLj and ADf and calculating the address separation E1, - a generator GF of the function f(E1), - a calculator CALVIT receiving the address ADLj sent by the said address determination circuit and determining the measured speed v= of the movable system as a function of the addresses ADL(nT + koT and ADL(nT) read at the sampling instants InT + koT) and nT, n and k0 being whole numbers, T being the time interval separating two successive sampling instants.

- an adder ADDIT effecting the sum -vm + f (E1) expressed in binary form, - a digital-analog converter receiving the signal delivered by the adder ADDIT and deiivering the sum (-v, + fIr1)) in the form of an analog signal.

- a device COMPRET for compensation of the average estimation delay of the measured speed vrn with respect to the speed v of the movable system supplying an analog signal F proportional to the measured acceleration',', - an adder ADD receiving the analog sum -vm + fIr1) and the signal rand supplying the set or instructed acceleration yc to the comparator.

5. Device according to claim 3 or 4, characterised in the calculator CALVIT comprises; - a circulating memory MEMOCIRC receiving the address ADL, sent by the circuit CIRCAD and conserving all the values of the address read at all the sampling instants comprised between the Instants (nT) and (nT + koT), - a subtractor-divider receiving the addresses ADL(nT) supplied by the circulating memory MEMOCIRC and ADL(nT + koT) sent by the circuit CIRCAD and calculating the measured speed v= equal to the difference between the addresses ADL(nT + koT) and ADL(nT) divided by the quantity koT, - a blocking device BLOC for blocking the value of the speed vrn during a time interval equal to the interval separating the two sampling Instants.

6. Device according to any of claims 3.4 or 5, characterised in that the addresses of the tracks being written on the carrier in a first binary code the circuit CIRCAD for determination of the address read comprises: - a threshold circuit GS transforming the series of analog pulses delivered by the read head into a series of logic pulses constituting the address ADG expresses in the said first code.

- a register-transcoder TRANSCOD transforming the address ADG1 into the address ADL1 expressed in a secondary binary code, - a sampling generator supplying sampling pulses permitting determination of the said sampling instants and controlling the register TRANSCOD in such a way that the latter delivers the address ADL at the same instants as the means ACCEL for calculation of the set or Instructed acceleration &gamma;c.

7. Method for displacement of a movable system with respect to a data carrier substantially as hereinbefore described with reference to the accompanying drawings.

8. Device for displacement of a movable system with respect to a data carrier substantially as hereinbefore described with reference to the accompanying drawings.