Method and apparatus for adaptive control

Claim 2

In a continuous analog signal control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method of adaptively calibrating a plant gain for regions of said disc and for each of said transducers, by determining a plant gain adjustment factor, said method comprising the steps of: commanding said actuator to a position r with a control signal u; generating a single-frequency sinusoid signal u s ; injecting said single-frequency sinusoid signal u s to said control system so that said single-frequency sinusoid is added to said control signal; combining said control signal u and said single-frequency sinusoidal signal u s to generate a corrected control signal u c ; multiplying said control signal and said added sinusoid signal by said sinusoid signal to generate a first product signal; measuring a position of said transducer y m ; multiplying said measured transducer position by said sinusoid value to generate a second product signal; low-pass filtering said first product signal to generate a first filtered signal; low-pass filtering said second product signal to generate a second filtered signal; dividing said second filtered signal by said first filtered signal to generate an identification ratio Q proportional to the measured plant gain; determining a new calibrated plant gain adjustment factor k g , said factor being an inverse function of said identification ratio Q; replacing a previous plant gain adjustment factor by said new calibrated plant gain adjustment factor; multiplying said control signal u c by said new calibrated plant gain adjustment factor k g to obtain an adjusted control signal u s ; wherein said control system automatically compensates for plant bias forces and for changes in effective plant gain.

Claim 3

The method as in claim 1, wherein said step of determining said new calibrated plant gain adjustment factor comprises dividing said identification ratio Q by a nominal identification ratio Q nominal to determine said new calibrated plant gain adjustment factor k g .

Claim 4

The method as in claim 1, wherein said single-frequency sinusoid is added directly to said control signal u at a control node.

Claim 5

The method as in claim 1, wherein said single frequency sinusoid is of the form u s =A 1 sin(wt).

Claim 6

The method as in claim 1, wherein said method includes measuring the plant phase by phase shifting, said phase shifting comprising the steps of: phase shifting said plant control signal u c , by a first phase shift Δp 1 so that said identification ratio Q is zero; phase shifting said transducer position y m by a second phase shift Δp 2 so that the reciprocal of said identification ratio 1/Q is zero; computing said plant phase from the difference of said first and second phase shifts.

Claim 7

The method of claim 1, wherein said method additionally comprises performing ROM and RAM calibration, said calibration comprising the steps of: measuring the A/D offset in high and low gain modes; measuring the power amplifier offset in both high and low gain modes; measuring the demodulator offset; measuring the plant gain in track following mode using a time averaging technique; and compensating each of the theoretical values used in said estimator by said measured values; whereby the accuracy of said plant gain is improved.

Claim 8

The method of claim 5, wherein said method of adaptively calibrating a plant gain includes a method of measuring a plant resonance frequency, said method of measuring a plant resonance frequency comprising: varying the frequency of said single-frequency sinusoid over a frequency range; and sampling said plant gain as said frequency is varied at a sampling rate sufficient to detect said resonance frequency.

Claim 9

The method of claim 7, wherein said method additionally comprises canceling the effect of a resonance frequency, wherein said method of canceling said resonance effects comprises filtering said resonance frequency.

Claim 10

The method of claim 8, wherein said step of filtering said resonance frequency comprises filtering said resonance frequency using an adaptive notch filter.

Claim 11

The method of claim 9, wherein said notch filter is a digital filter notch filter.

Claim 12

The method as in claim 10, wherein said digital notch filter is implemented in a look-up table.

Claim 13

The method as in claim 10, wherein said digital filter is a function of said resonance frequency.

Claim 14

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, a control system comprising an estimator, a circuit for detecting pulse signals encoding coarse position information as current track number T cn and demodulating an analog signal to obtain a position error signal related to the fine relative position of said transducer with respect to the center of said track, said system providing said current track number T cn , and said position error signal; a method of linearizing the positioning of said transducer over a seek length spanning a plurality of tracks, said method comprising: generating a precision linear position output value y m over said seek length by combining said current track number T cn value, and said position error signal value.

Claim 15

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, a control system comprising an estimator, a circuit for detecting pulse signals encoding coarse position information as current track number T cn and demodulating an analog signal to obtain a position error signal related to the fine relative position of said transducer with respect to the center of said track, said system providing said current track number T cn , and said position error signal; a method of linearizing the positioning of said transducer over a seek length spanning a plurality of tracks, said method comprising: generating a precision linear position output value y m over said seek length by combining said current track number T cn value, and said position error signal value.

Claim 16

The method of claim 13, wherein said generating a precision linear position output value y m , from said current track number T cn , and said position error signal value is according to the relations, y.sub.m =(2.sup.n ×T.sub.cn)+(k.sub.norm ×(AD.sub.mv -PES))+AD.sub.mv, for T.sub.cn an even number, and y.sub.m =(2.sup.n ×T.sub.cn)-(k.sub.norm ×(AD.sub.mv -PES))+AD.sub.mv, for T cn an odd number, where n is said A/D convertor word bit length, k norm is a multiplicative range normalizing constant, PES is the digitized PES value, T cn is the current track number, and AD mv is the mid-point value of the digital output of the A/D converter.

Claim 17

The method as in claim 14, wherein said current track number is encoded using a modulo-74 Grey code.

Claim 18

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method for maintaining said transducer in a centered position relative to said track, said method comprising the steps of: reading servo data from said disc; calculating a position of said transducer from said servo position data; calculating a predicted position, velocity, and bias, from previous predicted position, velocity, bias, and control values, and current control value; generating a control value based on said predicted position, velocity, and bias values; limiting said control value to a Predetermined maximum value; multiplying said control value by a plant gain adjustment factor k g to generate a compensated control value; shifting the level of said compensated control value to a Predetermined range; generating a compensated control signal in response to said compensated control value; applying said control signal to an amplifier to drive said actuator; calculating an estimator error e between said measured transducer position y m and said Predicted transducer position; pre-calculating new predicted states for position, velocity, and bias, based on said previous predicted position, velocity, and bias, said estimator error, and said current and previous control values; updating said control value by replacing the previous control value with the present control value; repeating said method for each actuator servo interrupt.

Claim 19

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method for maintaining said transducer in a centered position relative to said track, said method comprising the steps of: reading servo data from said disc; calculating a position of said transducer from said servo position data; calculating a predicted position, velocity, and bias, from previous predicted position, velocity, bias, and control values, and current control value; generating a control value based on said predicted position, velocity, and bias values; limiting said control value to a Predetermined maximum value; multiplying said control value by a plant gain adjustment factor k g to generate a compensated control value; shifting the level of said compensated control value to a Predetermined range; generating a compensated control signal in response to said compensated control value; applying said control signal to an amplifier to drive said actuator; calculating an estimator error e between said measured transducer position y m and said Predicted transducer position; pre-calculating new predicted states for position, velocity, and bias, based on said previous predicted position, velocity, and bias, said estimator error, and said current and previous control values; updating said control value by replacing the previous control value with the present control value; repeating said method for each actuator servo interrupt.

Claim 20

The method in claim 16, wherein said step of generating a control value based on said Predicted position, velocity, and bias values is accomplished according to the relation u=a 1 ×y m +pre - result, where a 1 =-K L, pre - result=(K×L×C e -K)×x+K(1)×r, K=control gain, L=estimator gain, and C e is the output vector that translates the estimated state to the measurement.

Claim 21

The method in claim 16, wherein said step of calculating new predicted states is accomplished according to the relations, x.sub.1 =x.sub.1 +b.sub.1 ×x.sub.2 +b.sub.2 ×x.sub.3 +b.sub.3 ×u+b.sub.4 ×u.sub.k-1 +b.sub.5 ×e, x.sub.2 =x.sub.2 +b.sub.6 ×x.sub.3 +b.sub.7 ×u+b.sub.8 ×u.sub.k-1 +b.sub.9 ×e, and x.sub.3 =x.sub.3 +b.sub.10 ×e, where x 1 =predicted position, x 2 =predicted velocity, x 3 =predicted bias, e is the estimator error defined above, u is the present control, u k-1 is the previous control, b 1 =A e (1,2), b 2 =A e (1,3), b 3 =B e1 (1), b 4 =B e2 (1), b 5 =L(1)+A e (1,2)×L(2)+A e (1,3)×L(3), b 6 =A e (2,3), b 7 =B e1 (1), b 8 =B e2 (2), b 9 =L(2)+A e (2,3)×L(3), and b 10 =L(3).

Claim 22

The method of claim 17, wherein said method additionally comprises the step of pre-calculating said pre - result value after said step of updating said control value so that said pre - result value is available for a subsequent execution of said method.

Claim 23

The method of claim 16, wherein said method additionally comprises the step of error checking.

Claim 24

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method for maintaining said transducer in a centered position relative to said track, said method comprising the steps of: reading servo data from said disc; calculating a position of said transducer from said servo position data; calculating the estimator error e, according to the relations e=y m -C e ×x=y m -x 1 , where x 1 =predicted position, x 2 =predicted velocity, and x 3 =predicted bias. calculating a predicted position, velocity, and bias, from previous predicted position, velocity, bias, and control values, and current control value; generating a control value based on said predicted position, velocity, and bias values; limiting said control value to a predetermined maximum value; multiplying said control value by a plant gain adjustment factor k g to generate a compensated control value; shifting the level of said compensated control value to a predetermined range; generating a compensated control signal in response to said compensated control value; applying said control signal to an amplifier to drive said actuator; pre-calculating new predicted states for position, velocity, and bias, based on said previous predicted position, velocity, and bias, said estimator error, and said current and previous control values; updating said control value by replacing the previous control value with the present control value; repeating said method for each actuator servo interrupt.

Claim 25

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method for maintaining said transducer in a centered position relative to said track, said method comprising the steps of: reading servo data from said disc; calculating a position of said transducer from said servo position data; calculating the estimator error e, according to the relations e=y m -C e ×x=y m -x 1 , where x 1 =predicted position, x 2 =predicted velocity, and x 3 =predicted bias. calculating a predicted position, velocity, and bias, from previous predicted position, velocity, bias, and control values, and current control value; generating a control value based on said predicted position, velocity, and bias values; limiting said control value to a predetermined maximum value; multiplying said control value by a plant gain adjustment factor k g to generate a compensated control value; shifting the level of said compensated control value to a predetermined range; generating a compensated control signal in response to said compensated control value; applying said control signal to an amplifier to drive said actuator; pre-calculating new predicted states for position, velocity, and bias, based on said previous predicted position, velocity, and bias, said estimator error, and said current and previous control values; updating said control value by replacing the previous control value with the present control value; repeating said method for each actuator servo interrupt.

Claim 26

The method in claim 21, wherein said step of generating a control value based on said predicted position, velocity, and bias values is accomplished according to the relation u=a 1 ×e+pre - resultA, where a 1 =-K L, pre - resultA=-K(1)×(x 1 -r)-K(2)×x 2 -(1/k s3 )×x 3 , and k s3 =a scale factor for the bias estimate.

Claim 27

The method in claim 21, wherein said step of calculating new predicted states is accomplished according to the relations, x.sub.1 =x.sub.1 +b.sub.1 ×x.sub.2 +b.sub.2 ×x.sub.3 +b.sub.3 ×u+b.sub.4 ×u.sub.k-1 +b.sub.5 ×e, x.sub.2 =x.sub.2 +b.sub.6 ×x.sub.3 +b.sub.7 ×u+b.sub.8 ×u.sub.k-1 +b.sub.9 ×e,and x.sub.3 =x.sub.3 +b.sub.10 ×e, where x 1 =predicted position, x 2 =predicted velocity, x 3 =predicted bias, e is the estimator error defined above, u is the present control, u k-1 is the previous control, b 1 =A e (1,2), b 2 =A e (1,3), b 3 =B e1 (1), b 4 =B e2 (1), b 5 =L(1)+A e (1,2)×L(2)+A e (1,3)×L(3), b 6 =A e (2,3), b 7 =B e1 (1), b 8 =B e2 (2), b 9 =L(2)+A e (2,3)×L(3), and b 10 =L(3).

Claim 28

The method of claim 22, wherein said method additionally comprises the step of pre-calculating said pre - resultA value after said step of updating said control value so that said pre - resultA value is available for a subsequent execution of said method.

Claim 29

The method of claim 15, wherein said method additionally comprises the step of error checking.

Claim 30

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system; a method for maintaining said transducer in a centered position relative to said track, said method comprising the steps of: reading servo data from said disc; calculating a position of said transducer from said servo position data; generating a control value based on a transfer function depending on at least one prior transducer position ym and control value u; limiting said control value to a predetermined maximum value; multiplying said limited control value by a plant gain adjustment factor k g to generate a compensated control value; shifting the level of said compensated control value to a predetermined range; generating a compensated control signal in response to said compensated control value; applying said control signal to an amplifier to drive said actuator; updating said control value by replacing the previous control value with the present control value; repeating said method for each actuator servo interrupt.

Claim 31

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system; a method for maintaining said transducer in a centered position relative to said track, said method comprising the steps of: reading servo data from said disc; calculating a position of said transducer from said servo position data; generating a control value based on a transfer function depending on at least one prior transducer position ym and control value u; limiting said control value to a predetermined maximum value; multiplying said limited control value by a plant gain adjustment factor k g to generate a compensated control value; shifting the level of said compensated control value to a predetermined range; generating a compensated control signal in response to said compensated control value; applying said control signal to an amplifier to drive said actuator; updating said control value by replacing the previous control value with the present control value; repeating said method for each actuator servo interrupt.

Claim 32

The method in claim 26, wherein said step of generating a control value based on a transfer function depending on at least one prior transducer position ym and control value u, is in accordance with the relation u=c.sub.0 ×y.sub.m +pre.sub.- resultB, where pre - resultB is of the general form pre.sub.- resultB=c.sub.1 y.sub.m (k)+c.sub.2 y.sub.m (k-1)+. . . +c.sub.m y.sub.m (k-n+1)-d.sub.1 u(k)-d.sub.2 u(k-1)-. . . -d.sub.m u(k-n+1).

Claim 33

The method of claim 27, wherein said method additionally comprises the step of pre-calculating said pre - resultB value after said step of updating said control value so that said pre - resultB value is available for a subsequent execution of said method.

Claim 34

The method of claim 26, wherein said method additionally comprises the step of error checking.

Claim 35

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system; a method for switching among a plurality of amplifier gain modes, said method for switching comprising the steps of: determining a control value that will drive said actuator to a commanded position; switching said gain mode in response to the magnitude of the control value applied; adjusting said control system in response to said selected gain mode.

Claim 36

In a control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system; a method for switching among a plurality of amplifier gain modes, said method for switching comprising the steps of: determining a control value that will drive said actuator to a commanded position; switching said gain mode in response to the magnitude of the control value applied; adjusting said control system in response to said selected gain mode.

Claim 37

The method of claim 30, wherein said adjusting comprises altering control system scale factors for each state variable characterizing said control system, said actuator amplifier gains, and said feedback gains in response to said selected amplifier gain mode.

Claim 38

The method in claim 30, wherein said plurality of actuator drive modes consists of a high-gain mode and a low-gain mode.

Claim 39

The method in claim 32, wherein said actuator velocity command has the form f(y e )=(k 1tf /k 2tf )×(r-x 1 ), in said low-gain mode.

Claim 40

The method of claim 31, wherein said step of adjusting said control system scale factors for each state variable characterizing said control system, said actuator drive power amplifier gains, and said feedback gains in response to said selected drive mode is according to the relationships x.sub.2tf =(yscale.sub.tf /yscale.sub.sk)×(k.sub.s2tf /k.sub.s2sk)×x.sub.2sk x.sub.3tf =(k.sub.s1tf /k.sub.s1sk)×(k.sub.s3tf /k.sub.s3sk)×(k.sub.pask /k.sub.patf)×x.sub.3sk u.sub.tf =(k.sub.s1tf /k.sub.s1sk)×(k.sub.pask /k.sub.patf)×u.sub.sk.

Claim 41

The method of claim 30, wherein said method additionally comprises the step of adding a compensation value to said control for compensating said control for feedback gain discontinuity, said compensation value is being dependent on said drive mode, said feedback gains, and said scale factors.

Claim 42

The method of claim 34, wherein said method additionally comprises the step of adding a compensation value to said control for compensating said control for feedback gain discontinuity, said compensation value is being dependent on said drive mode, said feedback gains, and said scale factors.

Claim 43

In a control system for a system having an electromechanical actuator responsive to a command signal, at least one position sensor for sensing actuator positional data, said position sensor being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed positional data, and a control system comprising an estimator; a method of adaptively calibrating a plant gain for regions of said actuator movement and for each of said position sensors, by determining a plant gain adjustment factor, said method comprising the steps of: commanding said actuator to a position r with a control signal u; generating a single-frequency sinusoid signal u s ; injecting said single-frequency sinusoid signal u s to said control system so that said single-frequency sinusoid is added to said control signal; multiplying said control signal and said added sinusoid signal by said sinusoid signal to generate a first product signal; measuring a position of said actuator y m based on said position sensor; multiplying said measured transducer position by said sinusoid value to generate a second product signal; low-pass filtering said first product signal to generate a first filtered signal; low-pass filtering said second product signal to generate a second filtered signal; dividing said second filtered signal by said first filtered signal to generate an identification ratio Q proportional to the measured plant gain; dividing said identification ratio Q by a nominal identification ratio Q nominal to determine a new calibrated plant gain adjustment factor k g ; replacing a previous plant gain adjustment factor by said new calibrated plant gain adjustment factor; multiplying said control signal u by said new calibrated plant gain adjustment factor k g to obtain a compensated control signal u a ; whereby said control system is able to automatically compensate for plant bias forces as well as adapt to changes in effective plant gain.

Claim 44

In a control system for a system having an electromechanical actuator responsive to a command signal, at least one position sensor for sensing actuator positional data, said position sensor being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed positional data, and a control system comprising an estimator; a method of adaptively calibrating a plant gain for regions of said actuator movement and for each of said position sensors, by determining a plant gain adjustment factor, said method comprising the steps of: commanding said actuator to a position r with a control signal u; generating a single-frequency sinusoid signal u s ; injecting said single-frequency sinusoid signal u s to said control system so that said single-frequency sinusoid is added to said control signal; multiplying said control signal and said added sinusoid signal by said sinusoid signal to generate a first product signal; measuring a position of said actuator y m based on said position sensor; multiplying said measured transducer position by said sinusoid value to generate a second product signal; low-pass filtering said first product signal to generate a first filtered signal; low-pass filtering said second product signal to generate a second filtered signal; dividing said second filtered signal by said first filtered signal to generate an identification ratio Q proportional to the measured plant gain; dividing said identification ratio Q by a nominal identification ratio Q nominal to determine a new calibrated plant gain adjustment factor k g ; replacing a previous plant gain adjustment factor by said new calibrated plant gain adjustment factor; multiplying said control signal u by said new calibrated plant gain adjustment factor k g to obtain a compensated control signal u a ; whereby said control system is able to automatically compensate for plant bias forces as well as adapt to changes in effective plant gain.

Claim 45

The method as in claim 37, wherein said single-frequency sinusoid is added directly to said control signal u at a control node.

Claim 46

The method as in claim 37, wherein said control system is a sampled-data control system, and each of said signals is a sampled data signal.

Claim 47

A rotating disc storage apparatus having data stored in tracks thereon, said device comprising: means for transducing said stored data into a data signal; an electro-mechanical actuator movable in response to a control signal, said actuator being mechanically linked to said means for transducing so that said means for transducing is moved from one of said plurality of tracks to another of said plurality of tracks in response to said control signal; means for controlling said actuator, said means for controlling generating said control signal which characteristics depend on an adjustable plant gain; means for estimating an actuator position, a actuator velocity, and an actuator bias from a measured transducer position and a control signal; means for generating a sum signal from an addition of a single-frequency sinusoid to said control signal; means for modulating said sum signal with said single-frequency sinusoid to generate a modulated control signal; means for modulating said measured transducer position with said single-frequency sinusoid to generate a modulated position signal; a low-pass filter for extracting low-frequency components of said modulated control and said modulated position signals; means for comparing a ratio of said low-pass filtered modulated control signal to said low-pass filtered modulated position signal to generate a plant gain adjustment factor; means for adjusting a gain characteristic of said means for estimating and said means for controlling in response to said plant gain adjustment factor.

Claim 48

A rotating disc storage apparatus having data stored in tracks thereon, said device comprising: means for transducing said stored data into a data signal; an electro-mechanical actuator movable in response to a control signal, said actuator being mechanically linked to said means for transducing so that said means for transducing is moved from one of said plurality of tracks to another of said plurality of tracks in response to said control signal; means for controlling said actuator, said means for controlling generating said control signal which characteristics depend on an adjustable plant gain; means for estimating an actuator position, a actuator velocity, and an actuator bias from a measured transducer position and a control signal; means for generating a sum signal from an addition of a single-frequency sinusoid to said control signal; means for modulating said sum signal with said single-frequency sinusoid to generate a modulated control signal; means for modulating said measured transducer position with said single-frequency sinusoid to generate a modulated position signal; a low-pass filter for extracting low-frequency components of said modulated control and said modulated position signals; means for comparing a ratio of said low-pass filtered modulated control signal to said low-pass filtered modulated position signal to generate a plant gain adjustment factor; means for adjusting a gain characteristic of said means for estimating and said means for controlling in response to said plant gain adjustment factor.

Claim 51

In a sampled-data control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method of adaptively calibrating a plant gain for regions of said disc and for each of said transducers, by determining a plant gain adjustment factor, said method comprising the steps of: commanding said actuator to a position r with a control signal u; generating a single-frequency sinusoid signal u s ; injecting said single-frequency sinusoid signal u s to said control system so that said single-frequency sinusoid is added to said control signal; combining said control signal u and said single-frequency sinusoidal signal u s to generate a corrected control signal u c ; multiplying said control signal and said added sinusoid signal by said sinusoid signal to generate a first product signal; measuring a position of said transducer y m ; multiplying said measured transducer position by said sinusoid value to generate a second product signal; low-pass filtering said first product signal to generate a first filtered signal; low-pass filtering said second product signal to generate a second filtered signal; dividing said second filtered signal by said first filtered signal to generate an identification ratio Q proportional to the measured plant gain; determining a new calibrated plant gain adjustment factor k g , said factor being an inverse function of said identification ratio Q; replacing a previous plant gain adjustment factor by said new calibrated plant gain adjustment factor; multiplying said control signal u by said new calibrated plant gain adjustment factor k g to obtain an adjusted control signal u s ; wherein each of said signals is a sampled data signal; wherein said control system automatically compensates for plant bias forces and for changes in effective plant gain.

Claim 52

In a sampled-data control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method of adaptively calibrating a plant gain for regions of said disc and for each of said transducers, by determining a plant gain adjustment factor, said method comprising the steps of: commanding said actuator to a position r with a control signal u; generating a single-frequency sinusoid signal u s ; injecting said single-frequency sinusoid signal u s to said control system so that said single-frequency sinusoid is added to said control signal; combining said control signal u and said single-frequency sinusoidal signal u s to generate a corrected control signal u c ; multiplying said control signal and said added sinusoid signal by said sinusoid signal to generate a first product signal; measuring a position of said transducer y m ; multiplying said measured transducer position by said sinusoid value to generate a second product signal; low-pass filtering said first product signal to generate a first filtered signal; low-pass filtering said second product signal to generate a second filtered signal; dividing said second filtered signal by said first filtered signal to generate an identification ratio Q proportional to the measured plant gain; determining a new calibrated plant gain adjustment factor k g , said factor being an inverse function of said identification ratio Q; replacing a previous plant gain adjustment factor by said new calibrated plant gain adjustment factor; multiplying said control signal u by said new calibrated plant gain adjustment factor k g to obtain an adjusted control signal u s ; wherein each of said signals is a sampled data signal; wherein said control system automatically compensates for plant bias forces and for changes in effective plant gain.

Claim 53

The method as in claim 42, wherein said single-frequency sinusoid is of the form u s =A 1 sin(wkT).

Claim 54

The method as in claim 42, wherein a plurality of cycles of said signals are bypassed before sampling so that transients of the filter outputs dissipate prior to sampling.

Claim 55

The method as in claim 42, wherein said first and second filter signals are signal processed over a plurality of sine wave periods to yield a more precise measurement of the plant characteristic.

Claim 56

The method as in claim 45, wherein said signal processing comprises averaging each of said first and second filter signals over a plurality of sine wave periods so that any ripple in said filtered signals is removed. PG,80

Claim 57

The method as in claim 42, wherein said step of determining said new calibrated plant gain adjustment factor comprises dividing said identification ratio Q by a nominal identification ration Q nominal to determine said new calibrated plant gain adjustment factor k g .

Claim 58

The method as in claim 42, wherein said single-frequency sinusoid is added directly to said control signal u at a control node.

Claim 59

The method as in claim 42, wherein said method includes measuring the plant phase by phase shifting, said phase shifting comprising the steps of: phase shifting said plant control signal u c , by a first phase shift Δp 1 so that said identification ratio Q is zero; phase shifting said transducer position y m by a second phase shift Δp 2 so that the reciprocal of said identification ration 1/Q is zero; computing said plant phase from the difference of said first and second phase shifts.

Claim 60

The method of claim 42, wherein said method additionally comprises performing ROM and RAM calibration, said calibration comprising the steps of: measuring the A/D offset in high and low gain modes; measuring the power amplifier offset in both high and low gain modes; measuring the demodulator offset; measuring the plant gain in track following mode using a time averaging technique; and compensating each of the theoretical values used in said estimator by said measured values; whereby the accuracy of said plant gain is improved.

Claim 61

The method of claim 50, wherein said method of adaptively calibrating a plant gain includes a method of measuring a plant resonance frequency, said method of measuring a plant resonance frequency comprising: varying the frequency of said single-frequency sinusoid over a frequency range; and sampling said plant gain as said frequency is varied at a sampling rate sufficient to detect said resonance frequency.

Claim 62

The method of claim 51, wherein said method additionally comprises canceling the effect of a resonance frequency, wherein said method of canceling said resonance effects comprises filtering said resonance frequency.

Claim 63

The method of claim 52, wherein said step of filtering said resonance frequency comprises filtering said resonance frequency using an adaptive notch filter.

Claim 64

The method of claim 53, wherein said notch filter is a notch filter selected from the group consisting of (i) a digital filter notch filter, (ii) a digital notch filter implemented in a look-up table, and (iii) a digital notch filter that filters as a function of said resonance frequency.

Claim 65

In a hybrid analog and digital control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method of adaptively calibrating a plant gain for regions of said disc and for each of said transducers, by determining a plant gain adjustment factor, said method comprising the steps of: commanding said actuator to a position r with a control signal u; generating a single-frequency sinusoid signal u s ; injecting said single-frequency sinusoid signal u s to said control system so that said single-frequency sinusoid is added to said control signal; combining said control signal u and said single-frequency sinusoidal signal u s to generate a corrected control signal u c ; multiplying said control signal and said added sinusoid signal by said sinusoid signal to generate a first product signal; measuring a position of said transducer y m ; multiplying said measured transducer position by said sinusoid value to generate a second product signal; low-pass filtering said first product signal to generate a first filtered signal; low-pass filtering said second product signal to generate a second filtered signal; dividing said second filtered signal by said first filtered signal to generate an identification ratio Q proportional to the measured plant gain; determining a new calibrated plant gain adjustment factor k g , said factor being an inverse function of said identification ratio Q; replacing a previous plant gain adjustment factor by said new calibrated plant gain adjustment factor; multiplying said control signal u by said new calibrated plant gain adjustment factor k g to obtain an adjusted control signal u s ; wherein said control system automatically compensates for plant bias forces and for changes in effective plant gain.

Claim 66

In a hybrid analog and digital control system for a magnetic disc storage device comprising an electromechanical actuator responsive to a command signal, at least one magnetizable rotating disc having disc track position data encoded magnetically thereupon, at least one transducer proximate said rotating disc surface for sensing said magnetic positional data, said transducer being mechanically linked to said actuator and generating a signal related to a measured position y m in response to said sensed magnetic positional data, and a control system comprising an estimator; a method of adaptively calibrating a plant gain for regions of said disc and for each of said transducers, by determining a plant gain adjustment factor, said method comprising the steps of: commanding said actuator to a position r with a control signal u; generating a single-frequency sinusoid signal u s ; injecting said single-frequency sinusoid signal u s to said control system so that said single-frequency sinusoid is added to said control signal; combining said control signal u and said single-frequency sinusoidal signal u s to generate a corrected control signal u c ; multiplying said control signal and said added sinusoid signal by said sinusoid signal to generate a first product signal; measuring a position of said transducer y m ; multiplying said measured transducer position by said sinusoid value to generate a second product signal; low-pass filtering said first product signal to generate a first filtered signal; low-pass filtering said second product signal to generate a second filtered signal; dividing said second filtered signal by said first filtered signal to generate an identification ratio Q proportional to the measured plant gain; determining a new calibrated plant gain adjustment factor k g , said factor being an inverse function of said identification ratio Q; replacing a previous plant gain adjustment factor by said new calibrated plant gain adjustment factor; multiplying said control signal u by said new calibrated plant gain adjustment factor k g to obtain an adjusted control signal u s ; wherein said control system automatically compensates for plant bias forces and for changes in effective plant gain.

Claim 67

The method as in claim 55, wherein said step of determining said new calibrated plant gain adjustment factor comprises dividing said identification ratio Q by a nominal identification ration Q nominal to determine said new calibrated plant gain adjustment factor k g .

Claim 68

The method as in claim 55, wherein said single-frequency sinusoid is added directly to said control signal u at a control node.

Claim 69

The method as in claim 55, wherein said single-frequency sinusoid is of the form u s =A 1 sin(wkT).

Claim 70

The method as in claim 55, wherein said method includes measuring the plant phase by phase shifting, said phase shifting comprising the steps of: phase shifting said plant control signal u c , by a first phase shift Δp 1 so that said identification ratio Q is zero; phase shifting said transducer position y m by a second phase shift Δp 2 so that the reciprocal of said identification ration 1/Q is zero; computing said plant phase from the difference of said first and second phase shifts.

Claim 71

The method of claim 55, wherein said method additionally comprises performing ROM and RAM calibration, said calibration comprising the steps of: measuring the A/D offset in high and low gain modes; measuring the power amplifier offset in both high and low gain modes; measuring the demodulator offset; measuring the plant gain in track following mode using a time averaging technique; and compensating each of the theoretical values used in said estimator by said measured values; whereby the accuracy of said plant gain is improved.

Claim 72

The method of claim 55, wherein said method of adaptively calibrating a plant gain includes a method of measuring a plant resonance frequency, said method of measuring a plant resonance frequency comprising: varying the frequency of said single-frequency sinusoid over a frequency range; and sampling said plant gain as said frequency is varied at a sampling rate sufficient to detect said resonance frequency.

Claim 73

The method of claim 55, wherein said method additionally comprises canceling the effect of a resonance frequency, wherein said method of canceling said resonance effects comprises filtering said resonance frequency.

Claim 74

The method of claim 62, wherein said step of filtering said resonance frequency comprises filtering said resonance frequency using an adaptive notch filter.

Claim 75

The method of claim 63, wherein said notch filter is a notch filter selected from the group consisting of (i) a digital filter notch filter, (ii) a digital notch filter implemented in a look-up table, and (iii) a digital notch filter that filters as a function of said resonance frequency.