Patent Application
20130070368
This application is based upon and claims the benefit of priority from the prior Japanese Patent Application No. 2011-205393, filed on Sep. 20, 2011, the entire contents of which are incorporated herein by reference.
Embodiments basically relate to a magnetic disk device and a head control method.
A head position error signal is used to control a position and a velocity of a head of a magnetic disk device. The head position error signal is obtained by reproducing servo information included in servo sectors on the surface of the magnetic disk. However, it is impossible to obtain the head position error signal after the head enters into a ramp mechanism during unload operation.
During the unload operation, the velocity of the head is estimated using a back electromotive force so that the velocity of the head can be controlled even within the ramp mechanism. The back electromotive force is generated by a voice control motor.
Aspects of this disclosure will become apparent upon reading the following detailed description and upon reference to the accompanying drawings.
As will be described below, in accordance with an embodiment, a magnetic disk device includes a head, a coil, a drive circuit, a detector, a first calculation unit, a generation unit, a first estimation unit, and a control unit. The head write in and read out information onto and from an information storage medium. The coil motor has a coil with terminals at both ends to move the head. The drive circuit drives the coil motor. The detector detects an inter-terminal voltage across the coil. The first calculation unit calculates a back electromotive force across the coil by the use of the inter-terminal voltage. The generation unit generates a target velocity as a reference value of a velocity of the head. The first estimation unit estimates a first velocity of the head and an error of the first velocity for each sampling time by the use of the back electromotive force. The first estimation unit estimates a second velocity for a succeeding sampling time immediately after the sampling time for decreasing the error. The control unit calculates a control instruction to bring the second velocity close to the target velocity.
In accordance with another embodiment, a controlling method of a head included in a magnetic disk device is described. The device includes a head, a coil, a drive circuit, a detector, a first calculation unit, a generation unit, a first estimation unit, and a control unit. The head write in and read out information onto and from an information storage medium. The coil motor has a coil with terminals at both ends to move the head. The drive circuit drives the coil motor. The detector detects an inter-terminal voltage across the coil. The first calculation unit calculates a back electromotive force across the coil by the use of the inter-terminal voltage. The generation unit generates a target velocity as a reference value of a velocity of the head. The first estimation unit estimates a first velocity of the head and an error of the first velocity by the use of the back electromotive force. The first estimation unit estimates a second velocity for decreasing the error. The control unit calculates a control instruction to bring the second velocity close to the target velocity.
The method includes the following steps:
detecting the inter-terminal voltage across the coil;
calculating the back electromotive force across the coil, the first calculation unit using the inter-terminal voltage to calculate the back electromotive force;
generating the target velocity, the generation unit generating the target velocity, the target velocity serving as the reference value of the moving velocity of the head;
estimating the first velocity and the error, the first estimation unit using the back electromotive force to estimate the first velocity and the error;
estimating the second velocity, the first estimation unit estimating the second velocity to reduce the error by; and
calculating the control instruction for bringing the second velocity close to the target velocity, the control unit calculating the control instruction.
Embodiments will be described below.
The magnetic disk device of
The head 1 is supported at an end of the arm 2. When the VCM 4 rotates the arm 2 around a rotation axis 3, the head 1 moves in a radius direction above a surface of the magnetic disk 5 rotated by a spindle motor (not shown), thereby performing seek operation and follow operation.
The head 1 can write in and read out information to and from the magnetic disk 5 at any given location. In the normal seek operation and follow operation, the MPU 10 uses a position error signal obtained from the servo information to perform the positioning control of the head 1.
When a shock detection sensor (not shown) detects a shock applied to the magnetic disk device, or when a user turns off the device, for example, the MPU 10 switches the control system from the above-described positioning control to the velocity control, thereby performing unload operation in which the head 1 is retracted to a ramp mechanism 6.
When a recovery to load operation from the unload-state is instructed, or when a user turns on the device, the MPU 10 performs the velocity control of the head 1 in place of the positioning control for the seek operation and the follow operation to move the head 1 from the ramp mechanism 6 to above the magnetic disk 5.
The VCM 4 is a coil motor provided with a magnet and a coil which are arranged to face each other, for example. The magnet is fixed to a base. The coil is provided to the axially-supported arm 2. When a current is passed through the coil, the VCM serves as an actuator which applies rotational force to the arm 2, i.e., drives the arm 2.
The VCM 4 can be shown using the equivalent circuit diagram as shown in
A VCM drive circuit 7 receives an instruction voltage from the MPU 10 in any cases of the positioning control and velocity control to pass the coil current Ivcm through the coil, thereby driving the arm 2.
The VCM drive circuit 7 includes a current feedback circuit. The VCM drive circuit 7 is separated from the VCM 4. In fact, the VCM drive circuit 7 and the VCM 4 are connected to each other. Accordingly, the coil of the VCM 4 may be included in a portion of the VCM drive circuit 7.
A detector 8 detects the coil inter-terminal voltage Vmeas. In the embodiment, the detector 8 is mentioned as a unit. Alternatively, the detector 8 may be provided as a portion of the VCM drive circuit 7.
A system configuration and operation of the MPU 10 of the magnetic disk device in accordance with the embodiment will be described with reference to
The MPU 10 is provided with a generation unit 20 to generate a targeted velocity of the head 1 (hereinafter referred to as a target velocity), a back electromotive force calculation unit 30 to calculate the back electromotive force across the coil, a disturbance estimation unit 40 to estimate disturbance in velocity applied to the head 1, a velocity estimation unit to estimate the velocity of the head 1, and a velocity control unit to control the velocity of the head 1, as modules.
The generation unit 20 generates the target velocity of the head 1 for each control cycle to input the target velocity into the velocity control unit 60 described later. The target velocity during the unload operation can be previously stored in the memory 9.
In
The back electromotive force calculation unit 30 calculates the back electromotive force Vbemf of the coil from the coil inter-terminal voltage Vmeas and the coil resistance value Rvcm for each control cycle.
The disturbance estimation unit 40 estimates the disturbance applied to the head 1 for each control cycle to calculate a parameter Rk. The parameter Rk adjusts a control bandwidth in the velocity control described later for each control cycle. The parameter Rk also shows how much the disturbance included in the back electromotive force is taken into consideration on estimating the velocity of the head 1.
The velocity estimation unit 50 estimates a state variable using the back electromotive force Vbemf and the above-described parameter Rk to minimize an error of the state variable for each control cycle. Here, the state variable is expressed by a vector including the velocity of the head 1 and the disturbance applied to the head 1 as its components.
The velocity control unit 60 uses the state variable estimated by the velocity estimation unit 50 to calculate a control instruction uk for each control cycle, thereby bringing the velocity of the head 1 close to the target velocity thereof generated by the target value generation unit 20.
The respective modules will be described in detail with reference to the block diagram for describing an operation of the MPU 10 shown in
The back electromotive force Vbemf is expressed by the following equation when the control cycle takes a time interval sufficient for attenuating the voltage caused by the inductance to remove effect of the inductance term:
V
bemf
=V
meas
−R
vcm
·I
vcm [Equation 1]
In the current feedback circuit, the coil current Ivcm is proportional to the instruction voltage Vvcm under the condition in which the current feedback sufficiently effects. The relationship therebetween can be expressed as Ivcm=βVvcm using a proportionality coefficient β, thereby enabling it to transform the back electromotive force Vbemf into the following equation:
In the above equation, “the proportionality coefficient β”דthe coil resistance Rvcm” is replaced by a and the value a can be regarded as a calculational coil resistance value.
In accordance with (Equation 2), the back electromotive force calculation unit 30 subtracts a multiplied value from the coil inter-terminal voltage Vmeas to calculate the back electromotive force Vbemf for each control cycle. Multiplying the coil resistance value a by the instruction voltage Vvcm provides the multiplied value. The coil inter-terminal voltage Vmeas is detected by the detector 8 to be outputted through an AD converter.
As shown in
The estimation unit 41 includes a disturbance observer of the background art, for example, to estimate the disturbance applied to the head 1. The estimation unit 41 uses the back electromotive force and the instruction voltage to calculate a disturbance d for each control cycle. The back electromotive force is calculated by the back electromotive force calculation unit 30.
The detector 42 uses the disturbance d calculated by the estimation unit 41 to detect the timing for the head 1 to run on the ramp mechanism 6. The detector 42 sequentially observes the disturbance d calculated for each control cycle. When the disturbance d becomes a value larger than a prescribed value to be predetermined, the detector 42 defines this time as start timing of running on by determining that the head 1 starts running on the ramp mechanism 6.
After the start timing, when the disturbance d becomes a value smaller than the prescribed, the detector 42 defines this time as end timing of running on by determining that the head 1 finishes running on the ramp mechanism 6.
The detector 42 sequentially calculates a time rate of change of the disturbance d. When this time rate of change becomes a value larger than a prescribed value to be predetermined, the detector 42 may determine that the head 1 starts running on the ramp mechanism 6. When the time rate of change becomes a value smaller than the prescribed value, the detector 42 may determine that the head 1 finishes running on the ramp mechanism 6.
The parameter calculation unit 43 obtains the start timing and the end timing from the detector 42. During the time interval from the start timing to the end timing, the parameter calculation unit 43 calculates the parameter Rk smaller than a parameter Rk for the routine operation. If the parameter Rk is small, a control bandwidth in the velocity control becomes high. If the parameter Rk is large, the control bandwidth becomes low. The routine operation excludes unload operation.
When the head 1 is running on the ramp mechanism 6, the parameter Rk is set small to heighten the control bandwidth in the velocity control, thereby allowing the head 1 to surely run on the ramp mechanism 6 without reducing the velocity of the head 1.
The parameter calculation unit 43 specifically passes the disturbance d calculated by the estimation unit 41 through a function f(d) to calculate the parameter Rk. The function f(d) has a hysteresis with respect to the change in the parameter Rk, for example.
As shown in
The back electromotive force across the coil is proportional to the velocity of the head 1, thereby performing the velocity control by using the back electromotive force during the unload operation of the head 1.
It is required for the head 1 to surely run on the ramp mechanism 6 in the velocity control. When the head 1 starts to run on the ramp mechanism 6, large external force acts on the head 1. The large external force greatly decreases the velocity of the head 1. Accordingly, the large external force possibly causes the head 1 to damage the surface of the magnetic disk 5 or to fail to run on the ramp mechanism 6. In order to heighten the control bandwidth in the velocity control, a control system is configured for a high gain.
However, the above-described back electromotive force includes noise. Accordingly, the control system is configured for a high gain so that the head 1 surely runs on the ramp mechanism 6. Such a control system also amplifies the noise to further cause audible noise.
In the embodiment, the velocity estimation unit 50 uses the time-varying Kalman filter to estimate the velocity of the head 1, thereby removing an influence of the noise included in the back electromotive force Vbemf. The Kalman filter is also made to be time varying so that a gain of the control system is suitable.
Specifically, an innovation gain M of the time-varying Kalman filter is set suitably, thereby estimating the state variable to minimize an error of mean square of the signal including noise. The back electromotive force Vbemf is used as an observed value.
When the control cycle is expressed by a sampling time k (=0, 1, 2, . . . ) below, an innovation gain Mk of the time-varying Kalman filter in the control cycle k is expressed by the following equation:
M
k
=
k
C
T(C
Pk is an error of the state variable. C is a coefficient matrix which relates the state of the system in the control cycle k to an observed value yk in the control cycle k.
A state variable xk can be estimated by the following equation using the innovation gain Mk and the observed value yk:
x
k
=
k
+M
k(yk−C
In the embodiment, the back electromotive force Vbemf can be used as the observed value yk. A vector including both the velocity estimate value and the disturbance estimate value can be used as a state variable xk.
An error Pk of the state variable xk can be updated by the following equation using the innovation gain Mk:
P
k=(I−MkC)
The state variable in the control cycle k+1, i.e., the succeeding sampling time immediately after each sampling time can be expressed by the following equation using the estimate value of the state variable obtained by the above-described (Equation 4):
k+1
=Ax
k
+B
i. [Equation 6]
uk is an input to a model of the VCM 4 in the control cycle k. A is a coefficient matrix to relate the state of the system in the control cycle k to the state of the system in the control cycle k+1. B is a coefficient matrix to relate the input uk in the control cycle k to the state of the system in the control cycle k+1.
An error Pk+1 of the state variable in the control cycle k+1, i.e., the succeeding sampling time immediately after each sampling time can be estimated by the following equation using the error Pk of the state variable obtained by the above-described (Equation 5):
k+1
=AP
k
A
T
+Q [Equation 7]
Q is a process noise and treated as a time-invariant parameter in the embodiment.
Accordingly, the above-described equations including (Equation 3) to (Equation 6) are calculated repeatedly, thereby enabling it to estimate the state variable given by (Equation 4) for each control cycle.
As the above-described coefficient matrixes A, B, and C, the same matrices as the respective coefficient matrices of an equation of state are used, for example. Prior inspection or the like provides the equation of state as a model expressing the characteristics of the VCM 4.
A gain updating unit 51 updates the innovation gain Mk of the time-varying Kalman filter for each control cycle. The gain updating unit 51 obtains the previously stored coefficient matrix C from the memory 9. The gain updating unit 51 updates the innovation gain Mk in accordance with the (Equation 3) using the coefficient matrix C, the parameter Rk, and a predicted value of a covariance matrix of an error. The parameter Rk is calculated by the parameter calculation unit 41. The predicted value of the covariance matrix of the error is calculated by a predicted value calculation unit 53 described later in the preceding sampling time immediately after each sampling time.
An estimate updating unit 52 calculates the state variable xk for each control cycle. The estimate updating unit 52 obtains the previously stored coefficient matrix C from the memory 9. The estimate updating unit 52 obtains the innovation gain Mk updated by the gain updating unit 51, the back electromotive force Vbemf calculated by the back electromotive force calculation unit 30, and a predicted value of the state variable. The predicted value of the state variable is calculated by the predicted value calculation unit 53 described later in the preceding sampling time immediately before each sampling time. The estimate updating unit 52 calculates the state variable xk in accordance with (Equation 4).
The estimate updating unit 52 uses the coefficient matrix C, the innovation gain Mk, and a predicted value of the covariance matrix of the error to calculate the covariance matrix Pk of the error for each control cycle in accordance with (Equation 5). The predicted value of the covariance matrix of the error is calculated by the predicted value calculation unit 53 described later in the preceding sampling time immediately before each sampling time.
The predicted value calculation unit 53 obtains each of the previously stored coefficient matrices A and B from the memory 9. The predicted value calculation unit 53 uses the state variable xk in the control cycle k and the control instruction uk in the control cycle k to calculate the predicted value of the state variable for the succeeding sampling time immediately after each sampling time in accordance with (Equation 6). The state variable xk in the control cycle k is updated by the estimate updating unit 52.
The predicted value calculation unit 53 obtains the coefficient matrix A and the previously stored process noise Q from the memory 9. The predicted value calculation unit 53 also obtains the covariance matrix Pk of the error in the control cycle k which is updated by the estimate updating unit 52. The predicted value calculation unit 53 calculates the predicted value of the covariance matrix Pk of the error for the succeeding sampling time immediately after each sampling time in accordance with (Equation 7).
As shown in
The difference-calculation unit 61 obtains the target velocity generated by the generation unit 20. The difference calculation unit 61 also obtains the state variable value xk to separate a component xvk of the velocity estimate value from the state variable xk. The state variable value xk is calculated by the estimate updating unit 52 for each control cycle. Multiplying the state variable xk by a matrix C1=[1 0] provides the component xvk, for example.
The difference calculation unit 61 subtracts the velocity estimate value xvk from the target velocity to calculate the velocity difference.
The control-instruction calculation unit 62 calculates a value as the control instruction uk. Multiplying the velocity difference by a gain K provides the value. The memory 9 stores the gain K previously. The difference calculation unit 61 calculates the velocity difference. The gain K can be previously obtained using a method of the background art from a parameter. The parameter is to determine the control characteristics such as a quick response or stability to be required at the time of the velocity control of the head 1, for example.
The disturbance-suppressing signal calculation unit 63 obtains the state variable xk calculated by the estimate updating unit 52 for each control cycle, separates a component xdk of a disturbance estimate value from the state variable xk. The disturbance-suppressing signal calculation unit 63 multiplies the component xdk by a minus to calculate a disturbance suppressing signal. The disturbance suppressing signal is to cancel out the disturbance which is estimated as the disturbance estimate value. Multiplying the state variable xk by a matrix C2=[0 −1] provides the disturbance suppressing signal.
When giving the control instruction uk to the VCM drive circuit 7, the drive-instruction calculation unit 64 calculates the drive instruction Vvcm which is actually to be given to the VCM drive circuit 7. The control instruction uk coincides with the drive instruction Vvcm in an ideal state. In fact, an amount of compensation due to external force is added to the drive instruction Vvcm as the disturbance suppressing signal in order to cancel out the influence of noise due to an external disturbance. The external force is to be experienced by the head 1 when the head 1 runs on the ramp mechanism 6.
The drive-instruction calculation unit 64 obtains the control instruction uk calculated by the control-instruction calculation unit 62 and a disturbance suppressing signal calculated by the disturbance-suppressing signal calculation unit 63. The control instruction uk and the disturbance suppressing signal are added to calculate the drive instruction Vvcm.
The velocity control unit 60 gives the drive instruction Vvcm calculated by the drive-instruction calculation unit 64 to the VCM drive circuit 7 for each control cycle and makes the velocity follow the target velocity to move the head 1.
In the embodiment, the time-varying Kalman filter eliminates the influence of noise when the velocity estimation unit 50 calculates the velocity estimate value of the head 1. The velocity control unit 60 uses the above-described velocity estimate value to calculate the control instruction, thereby enabling it to reduce audible noise during the unload operation. Hence, the embodiment can improve silence of the magnetic disk device during the unload operation.
The MPU 10 of the embodiment reduces audible noise more greatly than the MPU of the background art between the time of running on the ramp mechanism and the time of arriving at the stopper.
In the modification, the generation unit 20 uses the velocity estimate value xvk to generate a position target value of the head 1. The velocity estimate value xvk is calculated by the difference calculation unit 61 of the velocity control unit 60.
The generation unit 20 will be described in detail below. The same reference numerals will be used to denote the same or like portions throughout the figures below. Therefore, the same explanation will not be repeated.
The generation unit 20 generates the position target value of the head 1 for each control cycle. The memory 9 can store the position target value of the head 1 previously.
The generation unit 20 obtains the velocity estimate value xvk calculated by the difference calculation unit 61 and integrates the velocity estimate value xvk to calculate a position estimate value xrk as an estimate value of the position of the head 1 in each control cycle.
The generation unit 20 subtracts the position estimate value xrk from the above-described position target value to calculate a position difference.
The generation unit 30 multiplies the above-described position difference by a constant a, for example, to calculate the target velocity. The target velocity is inputted to the velocity control unit 60.
As a result, the start timing of running on the ramp mechanism 6 and the timing of colliding with the stopper can be reflected in more detail for the velocity control of the head 1, thereby enabling rapid unload operation.
The MPU of the background art hides the right timing of running on the ramp mechanism from the view of
As can be seen in
At least one of the above-described embodiments enables it to enhance silence of the magnetic disk device during the unload operation.
While certain embodiments have been described, those embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel embodiments described herein may be embodied in a variety of other forms; furthermore, various omissions, substitutions and changes in the form of the embodiments described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2011-205393 | Sep 2011 | JP | national |
