microactuator for high-capacity hard-disk drivesmaecourses.ucsd.edu/~pbandaru/mae268-sp09/class...
TRANSCRIPT
Page 1Hitachi Confidential2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Toshiki Hirano, San Jose Research Center
Microactuator for High-Capacity Hard-Disk Drives
Toshiki HiranoHitachi Global Storage Technologies
San Jose Research Center
Page 2Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Disk Drives in the past(IBM, 60 GB storage, 1980-1989)
Today:400 GB Hitachi Deskstar 7K400
Page 3Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Recording Density GrowthRecording al
Production Year Ed Grochowski
60 70 80 90 100 1101E-3
1E-2
1E-1
1E+0
1E+1
1E+2
1E+3
1E+4
1E+5
1E+6
Are
alD
ensi
ty M
egab
its/in
2
HGST Disk Drive ProductsIndustry Lab DemosHGST Disk Drives w/AFCDemos w/AFC
HGST Areal Density Perspective
1st MR Head
1st GMR Head
2000 10
60% CGR
Ultrastar 146Z10
Deskstar 180GXP
100% CGR
Corsair
Deskstar 16GP
Travelstar 30GNMicrodrive II
1st AFC Media
Future Are DensityProgress
Travelstar 80GN
Perpendicular
Superparamagnetic effect
105
104
103
102
106
10
1
10-1
10-2
10-3
25% CGR
IBM RAMAC (First Hard Disk Drive)
1st Thin Film Head3375
35 Million XIncrease
Page 4Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Track Density and Linear Density
1E+1
1E+2
1E+3
1E+4
1E+5
1E+6
90 92 94 96 98 100 102 104
Availability Year
Areal Density, mbits/in2
Track, Areal, Linear Density Perspective
02 04
10 5
10 4
10 3
10 2
10
Linear Density, kbits/in
Track Density, tracks/in
2000
Ed Grochowski
Track Density CGR = 50%
Areal Density CGR = 100%
Linear Density CGR = 30%
Circles = Server productsSquares = Mobile products
Hard Disk Drive Products
106
Linear Density
Trac
k D
ensi
ty
100 kTPI(250 nm pitch)
Page 5Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Track Pitch Challenge
1000X
95 mm95 m
Data track width:Data track width:250 nm
250 micrometer(human hair diameter x 3)
Page 6Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Minimizing Tracking Error
• The position error between magnetic head and data track on disk (called Track Mis-Registration-TMR) must be ~10% of track pitch.– For 250 nm track pitch, acceptable TMR is about 25 nm.
• Two factors affecting TMR:– Position Disturbance– Disturbance correction by close-loop control (servo)
• First-order TMR model:– TMR = Position Disturbance / Disturbance Correction by servo
Page 7Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Position Disturbance Sources
• Majority of position disturbance is caused by air-flow induced by spinning disk.– Example: 84 mm diameter disk spinning at 10,000 RPM– Linear speed at disk edge = 44 m/Sec ~= 100 MPH !
• Today’s disk drive has:– Extremely high stiffness (Disk, Actuator).– Aerodynamically designed mechanical components.– Optimization such that the vibration node coincides magnetic head
location.• Example: Suspension beam
• However, there is usually a limit on how well one can reduce this disturbance.– Example: Disk can not be infinitely thick, mechanical tolerances.
Page 8Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Disturbance rejection by servo
• Disk drive has a close-loop control system (servo) that corrects position disturbance.
• Close-loop consists of:– Position error measured by head -> Controller generates input to
voice-coil-motor (VCM) actuator -> Position error is corrected.
ServoController Head Position
Voice-Coil Motor Drive Signal
Page 9Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Servo Bandwidth• Servo performance is measured by “servo bandwidth”.
– Servo bandwidth is defined as the frequency where the open-loop gain crosses 0 dB line.
– Higher bandwidth enables better disturbance correction.– Typical servo bandwidth of current HDD product is 1-1.5 kHz– Higher bandwidth is required for future HDDs.
Gain
G(s)+
-+Input
Disturbance
Output(Position)
Error Rejection= = Position
Disturbance
1
1+G(s)
G(s): Controller x ActuatorFrequency(Hz)
0 dB
Servo Bandwidth
G(s)
Bode Plot of G(s)(Open-loop transfer function)
Page 10Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Servo bandwidth limitation• Servo bandwidth of current HDD is usually limited by mechanical
resonances of head-positioning actuator (VCM, arm, and suspension).
Pivot Friction
1/s^2(40 dB/dec)
Page 11Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Dual-Stage Actuator
• 1st Generation (Moving Suspension)– Prototyped by HTI, NHK, Magnecom, Fujitsu ...– PZT controlled– Limitation by suspension dynamics
• 2nd Generation (Moving Slider)– TDK PZT based (fragile)– MEMS prototyping: IBM/Hitachi, Seagate... – Challenge for fabrication / assembly
• 3rd Generation (Moving Magnetic Head)– Concept phase: U-Tokyo/Hitachi ...– Complicated integration of MEMS & magnetic
element
HTI
TDK
Hitachi
U-tokyo/Hitachi
Page 12Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
MEMS Microactuator Developed at Hitachi SJRC
• Rotary actuator• Electrostatically driven• Spring support (no friction)• Stroke: +- 1 mm• Actuator mass: 3-4 mg• Plated nickel structure• Output force: ~0.5 mN
Page 13Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Electrostatic Actuator• Electrostatic force:
Ng
hVF ×=2ε
• To have large force output:• large h• small g• high V• More N
• Process capability limits aspect ratio (h/g).
• High-Aspect-Ratio process was developed.
Page 14Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
High-Aspect Ratio Process• High-Aspect Ratio (up
to 20:1, or 40 um thick, 2 um wide)
• Metal (Nickel) Structure
• Process Steps:– Thick polymer coating– Highly directional oxygen
plasma etching of polymer (electro-plating mold).
– Nickel Electroplating.
2 um
40 um
Page 15Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Microactuator Design Example• Push-Pull• Three connections (Drive+, Drive-, GND or DC Bias)
2.1mm
1.7m
m
Anchor to the substrate(Fixed)
Slider Binding Platform(Movable)
Electrostatic Actuator
Rotational Spring
Page 16Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Fabrication
Metal
Anchor
Structure(40 um thick)
Stud(10 um)
Top cover(10 um)
• Mostly made by typical IC processes, such as photo-lithography, film deposition and etching.
• Some non-IC but magnetic head standard processes, such as electroplating
• Batch fabricated. – ~1500 chips from 4" wafer.– ~7000 chips from 8" wafer.
Page 17Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Process Steps (Cross Section)
Lithography and 2nd Ni plating
Polymer Coating (40 um)
Insulator coatingMetal Patterning
Lithography and 3rd Ni plating/Au plating
2nd Insulator PatterningSacrificial Patterning
High Aspect RatioPolymer Etching (20:1)
Thick nickel plating (40 um)
3rd Insulator Patterning
Polymer/resist removalReleaseBack-side lapping
Page 18Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Microactuator Chip Example
Page 19Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Example of finished wafer
8 inch~7000 Chips / wafer
Page 20Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Microactuator Assembly with slider and suspension
• Additional lead wire required:– Current 4 wires -> 7 wires
• Additional assembly step required:– Current two parts -> three parts
Page 21Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Microactuator assembly into drive
• Three extra wires
Page 22Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Microactuator Performance• Frequency response (From input voltage to position)
– Microactuator vs. conventional VCM actuator
-60-40-20
020406080
100 1000 10000 100000
Frequency (Hz)
Gai
n (d
B)
VCM Microactuator
-180
-90
0
90
180
100 1000 10000 100000
Frequency (Hz)
Pha
se (d
egre
e)Microactuator Mass/spring mode. Damped by close-loop controller
Page 23Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Servo Experiment
• Parallel design– Low frequency signal goes to VCM, and high frequency signal goes
to microactuator.
MicroactuatorController
VCMController
Microactuator
VCM
++
-
Digital Signal Processor Hard disk drive
MicroactuatorDisplacement
VCMDisplacement
PositionReferencePosition
PositionError
Page 24Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Servo Experiment Results (1)• Open-loop transfer function
– 0 dB cross-over frequency (servo bandwidth) of 8 kHz by microactuator achieved, compare to ~1.8 kHz conventional VCM servo.
-60
-40
-20
0
20
40
100 1000 10000 100000
Frequency (Hz)
Gai
n (d
B)
Microactuator+VCM VCM only
-180
-90
0
90
180
100 1000 10000 100000
Frequency (Hz)
Pha
se (d
egre
e)
Microactuator0 dB@8 kHz
VCM0 [email protected] kHz
Page 25Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Servo Experiment Results (2)• Error rejection function
– With microactuator, error rejection is better up to 6 kHz.
-50
-40
-30
-20
-10
0
10
100 1000 10000 100000
Frequency (Hz)
Erro
r Rej
ectio
n (d
B)
Microactuator+VCM VCM only
Page 26Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Contact recording with microactuator
– Microactuator has small moving mass, thus more sensitive to any external force.
10 100 1000 10000 100000
Frequency (Hz)
1.0E-12
1.0E-11
1.0E-10
1.0E-9
1.0E-8
1.0E-7
1.0E-6
Disturbance (m)
VCM Milli/VCM Micro/VCM
Force Input: 0.7mN Normal Contact Force * 0.3 Friction Coefficient * Sin(15) SkewVCM: Iz=4E-6 kg*m2, fr=70, Q=5, L=50mm, Milli: Iz= 4.5E-9, fr=9kHz, Q=30, L=11, Micro: Iz=5.7E-13, fr=2kHz, Q=1, L=0.625mm
Off-track Disturbance by Contact Force
Contact Force
Skew
k1
c1
Iz1 k2 c2
Iz2
Other MEMS Application in Data Storage
"MILLIPEDE""MILLIPEDE"
x
z 2
z3
z1y
Nanotechnology Nanotechnology entering Data Storage entering Data Storage
Thermomechanical Read/Write on Thin Polymer Media
32 x 32 (1024) Cantilever Array Chip
Micromechanical x|y|z Scanner
=>most recent areal density demo 1Tbits/inch2 (single lever)
40 nm pitch
40 nm bit size
Zurich Research LaboratoryMicro- and Nanomechanics Group
Courtesy of Dr. Peter Vettiger, IBM Zurich Lab
Page 28Toshiki Hirano, San Jose Research Center
2004/4/6Copyright 2003 Hitachi Global Storage Technologies
Conclusions• Moving-slider microactuator for tracking servo was
investigated.• Microactuator was successfully fabricated and
assembled into drive.• Servo experiment was carried out, and 8 kHz servo
bandwidth was demonstrated.– Conventional VCM has ~1.8 KHz bandwidth.
• Compatibility with contact recording needs to be investigated.
AcknowledgementThis research was carried out with other collaborators, including Henry Yang, Matthew White, Fu-Ying Huang, Francis Lee, Mr. Harpreet Singh, Ms. Karen Scott and other members at Hitachi San Jose Research Center.