1. Field of the Invention
The present invention relates to disk drives for computer systems. More particularly, the present invention relates to techniques for efficiently determining repeatable runout (RRO) in a disk drive.
2. Description of the Prior Art
With reference to
The RRO disturbance due to the disk having a non-centric alignment with the spindle motor is sinusoidal with a period equal to the rotation of the disk. This sinusoidal disturbance can be represented as:
a*cos(2πk/N)+b*sin(2πk/N)
where {a,b} are coefficients corresponding to the magnitude of the disturbance (magnitude of the non-centric offset) and k is an index representing one of N servo sectors.
Prior art techniques typically estimate the composite RRO disturbance due to all sources of eccentricity, such as eccentricities of the servo sectors (written-in RRO), spindle motor runout, disk warping, as well as the RRO disturbance due to the disk's non-centric alignment with the spindle motor. The prior art techniques for estimating the composite RRO disturbance typically involves processing the position error signal (PES) over many revolutions of the disk in order to average out other noise sources (the non-repeatable runout). The composite RRO disturbance is typically estimated during a manufacturing process, and may be updated every time the disk drive is powered on to account for changes that occur over time, particularly from disk slippage due to external physical shocks. However, the numerous revolutions needed for estimating the composite RRO disturbance may result in undesirable response delays at power-on.
There is, therefore, a need for a fast, efficient technique for estimating the RRO disturbance in a disk drive that may be subjected to a physical shock causing disk slippage.
The present invention may be embodied in a method for determining fundamental-frequency repeatable runout (1FRRO) coefficients in a disk drive. The disk drive may include a transducer head, a rotating magnetic disk having a plurality of concentric data tracks defined by embedded servo wedges that provide position information, and an actuator coupled to the head. In the method, the head is caused to move in a substantially constant velocity motion across a selected portion of the tracks of the rotating magnetic disk in response to a control signal. The position information is read from the embedded servo wedges as the head moves across of the selected portion of the tracks. The 1FRRO coefficients are determined based on the position information read as the head moved in a substantially constant velocity motion across the selected portion of the tracks.
In more detailed features of the invention, the step of causing the head to move in a substantially constant velocity motion may occur as part of a ramp-load operation. The position information read from the embedded servo wedges may be track identification values. Also, the 1FRRO coefficients may be for computing first feed-forward compensation values for use in a head-position control servo loop. The first feed-forward compensation values 1FRRO coefficients may be computed using the 1FRRO coefficients {a,b} of a sinusoid:
a*cos(2πk/N)+b*sin(2πk/N)
where k is an index representing one of N servo wedges.
In other more detailed features of the invention, the 1FRRO coefficients may comprise a first coefficient and a second coefficient. The 1FRRO coefficients {a,b} may be determined based on the following equations:
=[P(1), P(2), P(3), . . . , P(n)]T;
P(i)=TK(i+1)−TK(i)−v;
C(i)=cos(2*pi*W(i+1)/N)−cos(2*pi*W(i)/N);
S(i)=sin(2*pi*W(i+1)/N)−sin(2*pi*W(i)/N);
where:
v is an average velocity determined by an average number of tracks per sample;
i is the i-th sample of the position information read during the constant motion;
TK(i) is the track id reading at i-th sample;
N is the total number of servo wedges;
W(i) is a servo wedge number at sample i; and
n is the total number of samples of the position information.
The present invention also may be embodied in a disk drive including a transducer head, a rotating magnetic disk having a plurality of concentric data tracks defined by embedded servo wedges that provide position information, an actuator coupled to the head, and a control system. The control system may be operable to generate a control signal for causing the head to move in a substantially constant velocity motion across a select portion of the tracks of the rotating magnetic disk, and to determine fundamental-frequency repeatable runout 1FRRO coefficients for the disk drive based on the position information read as the head moved in a substantially constant velocity motion across the selected portion of the tracks.
Another embodiment of the invention may reside in a method for determining repeatable runout (mFRRO) coefficients for a higher harmonic of a fundamental-frequency in a disk drive. In the method, the head is caused to move in a substantially constant velocity motion across a selected portion of the tracks of the rotating magnetic disk in response to a control signal. The position information is read from the embedded servo wedges as the head moves across of the selected portion of the tracks. The mFRRO coefficients are determined based on the position information read as the head moved in a substantially constant velocity motion across the selected portion of the tracks.
In more detailed features of the invention, the step of causing the head to move in a substantially constant velocity motion may occur as part of a ramp-load operation. The position information read from the embedded servo wedges may be track identification values. The mFRRO coefficients may be for computing first feed-forward compensation values for use in a head-position control servo loop. The first feed-forward compensation values mFRRO coefficients may be computed using the mFRRO coefficients {a,b} of a sinusoid:
a*cos(2πmk/N)+b*sin(2πmk/N)
where k is an index representing one of N servo wedges, and
m is the harmonic of the fundamental frequency.
The mFRRO coefficients may comprise a first coefficient and a second coefficient. The mFRRO coefficients {a,b} may be determined based on the following equations:
=[P(1), P(2), P(3), . . . , P(n)]T;
P(i)=TK(i+1)−TK(i)−v;
C(i)=cos(2*pi*m*W(i+1)/N)−cos(2*pi*m*W(i)/N);
S(i)=sin(2*pi*m*W(i+1)/N)−sin(2*pi*m*W(i)/N);
where:
v is an average velocity determined by an average number of tracks per sample;
i is the i-th sample of the position information read during the constant motion;
TK(i) is the track id reading at i-th sample;
N is the total number of servo wedges;
W(i) is a servo wedge number at sample i;
n is the total number of samples of the position information, and
m is the harmonic of the fundamental frequency.
The accompanying drawings illustrate embodiments of the present invention and, together with the description, serve to explain the principles of the invention.
With reference to
The step of causing the head 32 to move in a substantially constant velocity motion may occur as part of a ramp-load operation where the head is load over the disk surface from a parking ramp 43. Smaller form-factor disk drives may be used in mobile devices that may be subject to shock forces. The shock forces may cause a mechanical slip in the position between a disk hub and the disk. The slip may cause a large 1FRRO that acts as a disturbance in the disk drive's servo system. Integrating the constant velocity sweep as part of a ramp-load operation further saves time because the ramp-load operation is a necessary part of disk-drive initialization at power-on.
Adaptive feed-forward cancellation (AFC) may used to cancel repeatable runout (RRO) in a disk drive 30 as shown with reference to the servo control loop 44 in
FFC(k)=A(k−1)cos(ωTk)+B(k−1)sin(ωTk)
The coefficients A and B are related to the magnitude and phase of the 1FRRO disturbance. The coefficients are learned each time the disk drive was powered up using the following equations.
A(k)=A(k−1)+λPES cos(ωTk)
B(k)=B(k−1)+λPES sin(ωTk)
where λ is a learning rate, which is set to a fast rate at power up and lowered to a slow rate during disk drive operation. However, prior to the present invention, the initial values of A and B were set to zero and the convergence rate of the coefficient learning could be as high as 30 disk revolutions.
This high number of disk revolutions often made it difficult to meet strict power-to-ready time specifications needed to provide an “instant-on” feel to mobile devices, which feel is also available in alternate competitive technologies such as flash memory. In the present invention, the 1FRRO coefficients may be used in computing first feed-forward compensation values for use in a head-position control servo loop. The first feed-forward compensation values 1FRRO coefficients may be computed using the 1FRRO coefficients {a,b} of a sinusoid:
a*cos(2πk/N)+b*sin(2πk/N)
where k is an index representing one of N servo wedges.
The 1FRRO coefficients {a,b} may be determined based on the following equations:
=[P(1), P(2), P(3), . . . , P(n)]T;
P(i)=TK(i+1)−TK(i)−v;
C(i)=cos(2*pi*W(i+1)/N)−cos(2*pi*W(i)/N);
S(i)=sin(2*pi*W(i+1)/N)−sin(2*pi*W(i)/N);
where:
v is an average velocity determined by an average number of tracks per sample;
i is the i-th sample of the position information read during the constant motion;
TK(i) is the track id reading at i-th sample;
N is the total number of servo wedges;
W(i) is a servo wedge number at sample i; and
n is the total number of samples of the position information.
The control system 47 of the disk drive 30 (
The magnetic media surface of the disk 34 is accessed using the head 32. The tracks 36 on the media surface may be divided into storage segments. Each storage segment typically begins with a servo sector which is followed by data sectors. The servo sector for a storage segment corresponds to an intersection with the radially-extending embedded servo wedges 38. Each servo sector includes a track address for generating a coarse position for the head, and servo bursts for generating a fine position of the head with respect to the centerline of the target track. The data sectors may include data blocks, each generally storing 512 data bytes. Each data block may be addressed using a logical block address (LBA).
With reference to
Although not shown in
With reference to
The mFRRO coefficients may be for computing first feed-forward compensation values for use in a head-position control servo loop 44. The first feed-forward compensation values mFRRO coefficients may be computed using the mFRRO coefficients {a,b} of a sinusoid:
a*cos(2πmk/N)+b*sin(2πmk/N)
where k is an index representing one of N servo wedges, and
m is the harmonic of the fundamental frequency.
The mFRRO coefficients may comprise a first coefficient and a second coefficient. The mFRRO coefficients {a,b} may be determined based on the following equations:
=[P(1), P(2), P(3), . . . , P(n)]T;
P(i)=TK(i+1)−TK(i)−v;
C(i)=cos(2*pi*m*W(i+1)/N)−cos(2*pi*m*W(i)/N);
S(i)=sin(2*pi*m*W(i+1)/N)−sin(2*pi*m*W(i)/N);
where:
v is an average velocity determined by an average number of tracks per sample;
i is the i-th sample of the position information read during the constant motion;
TK(i) is the track id reading at i-th sample;
N is the total number of servo wedges;
W(i) is a servo wedge number at sample i;
n is the total number of samples of the position information, and
m is the harmonic of the fundamental frequency.
A gain factor may need to be applied between the estimated a and b coefficients and the learned A and B coefficients. The gain factor may be calibrated and stored during manufacturing by learning calculating the coefficients A and B during manufacture. The a and b coefficients are estimated, and the gain factor calculated:
Gain=(a2+b2)1/2./(A2+B2)1/2
.
After power up and estimation the a and b 1FRRO coefficients, the initial coefficients A and B may be calculated by the following equations:
A=−a/Gain
B=−b/Gain
The resulting A and B values are used as the initial values of the AFC, and a much faster learning rate may be achieved as indicated in
Number | Name | Date | Kind |
---|---|---|---|
5550685 | Drouin | Aug 1996 | A |
5585976 | Pham | Dec 1996 | A |
5793559 | Shepherd et al. | Aug 1998 | A |
5825578 | Shrinkle et al. | Oct 1998 | A |
5920441 | Cunningham et al. | Jul 1999 | A |
6049440 | Shu | Apr 2000 | A |
6097565 | Sri-Jayantha et al. | Aug 2000 | A |
6141175 | Nazarian et al. | Oct 2000 | A |
6310742 | Nazarian et al. | Oct 2001 | B1 |
6545835 | Codilian et al. | Apr 2003 | B1 |
6654198 | Liu et al. | Nov 2003 | B2 |
6661599 | Chen et al. | Dec 2003 | B1 |
6738205 | Moran et al. | May 2004 | B1 |
6765747 | Sun et al. | Jul 2004 | B1 |
6847503 | Zhang et al. | Jan 2005 | B2 |
6952320 | Pollock et al. | Oct 2005 | B1 |
6999267 | Melkote et al. | Feb 2006 | B1 |
20040160696 | Meyer | Aug 2004 | A1 |
20040246619 | Zhang | Dec 2004 | A1 |
20050231842 | Pang et al. | Oct 2005 | A1 |