An optical head for a magneto-optic disk test system

Item Type text; Thesis-Reproduction (electronic)

Authors Bushroe, Frederick Nicholas, 1964-

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this materialis made possible by the University Libraries, University of Arizona.Further transmission, reproduction or presentation (such aspublic display or performance) of protected items is prohibitedexcept with permission of the author.

Download date 14/06/2021 15:18:24

Link to Item http://hdl.handle.net/10150/277154

http://hdl.handle.net/10150/277154

INFORMATION TO USERS

The most advanced technology has been used to photograph and reproduce this manuscript from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. These are also available as one exposure on a standard 35mm slide or as a 17" x 23" black and white photographic print for an additional charge.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University Microfilms International A Bell & Howell Information Company

300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600

Order Number 1339032

An optical head for a magneto-optic disk test system

BushToe, Frederick Nicholas,'M.S.

The University of Arizona, 1989

U M I 300 N. Zeeb Rd. Ann Arbor, MI 48106

AN OPTICAL HEAD FOR A MAGNETO-OPTIC

DISK TEST SYSTEM

by

Frederick Nicholas Bushroe

A Thesis Submitted to the Faculty of the

COMMITTEE ON OPTICAL SCIENCES (GRADUATE)

In Partial Fulfillment of the Requirements For the degree of

MASTER OF SCIENCE

In the Graduate College THE UNIVERSITY OF ARIZONA

1 9 8 9

2

STATEMENT BY AUTHOR

This thesis has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this thesis are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, rermission must be obtained from the author.

SIGNED:

APPROVAL BY THESIS DIRECTOR

This thesis has been approved on the date shown below:

J&nes J. Burke ^Professor of

Date Optical Sciences

3

ACKNOWLEDGMENTS

The optical head for the magneto-optic test system was made possible through the advice, cooperation, and efforts of many individuals. Scott DeVore deserves special thanks for his donated time, enthusiasm, and knowledgeable advice. Scott guided the head design from the beginning and worked with us through the alignment and fabrication processes.

I also thank my advisor Professor James Burke for his guidance, moral support, and faith in me, during my time at Optical Sciences. I am grateful to Tom Milster for managing our lab group and pulling the test system together in a short period of time.

I extend my gratitude to the following people and companies for their contributions listed below: Barbara Shula, Henryk Birecki, and John Graves of Hewlett-Packard Labs in Palo Alto, Ca. for the read/write electronics, equipment, and software; IBM, Siemens, and Eastman Kodak for sponsoring the Optical Data Storage Center; David Kay of Eastman Kodak, Glenn Sincerbox and Blair Finkelstein of IBM for their exchange of information on various design issues; Robert Uber for his practical advice and discussions on many aspects of the head; Kevin Erwin our technician for his valuable knowledge of optical and mechanical fabrication techniques; Mark Shi Wang for his willingness to learn the control software and his interfacing abilities; Fred Froehlich and Jim Mueller for their electronic design and fabrication; Bob Trusty for his characterization of the bias coil; Professor Roland Shack for his willingness to answer my questions on optical issues; Dan Vukobratovich for his time spent discussing mechanical issues related to the threaded collimating lens mount and the kinematic head mount; Charlie Brown and Derek Clark for their machine shop skills and cooperation; Tom Walker of OCLI for donation of the partial polarizing beamsplitter; and WYKO Corporation for the use of their Ladite instrument.

Aiding in the conceptualization of an optical head for a test system, were visits during the summer of 1988 to several companies in the optical storage arena. I am grateful to the following people for helping to arrange these visits: Scott Hamilton, Apex Systems Inc., Boulder, CO; Ron Brown, Applied Magnetics Optical Products Division, Monument, CO; Barbara Shula, Hewlett-Packard, Palo Alto, CA; Martin Chen, IBM Almaden, CA; Mr. O'Connor, Information Systems Inc., Colorado Springs, CO; Ralph Poplawsky, Laser Magnetic Storage, Colorado Springs, CO; Di Chen, Optotec, Colorado Springs, CO; Vincent Tarpey, Pencom, Sunnyvale, CA; and Tim Fayram, Verbatim, Sunnyvale, CA.

I would also like to thank our secretary Geri Simons for her help and the use of her laser printer. Finally, I wish to thank my mother, father, and family for their encouragement and support throughout my college education.

4

TABLE OF CONTENTS

LIST OF ILLUSTRATIONS 5

LIST OF TABLES 6

ABSTRACT 7

INTRODUCTION 8

OPERATING PRINCIPLES OF THE MAGNETO-OPTIC HEAD 9

DESIGN OF THE MAGNETO-OPTIC HEAD 20

Design Tradeoffs and Alternate Configurations 20

The Semiconductor Laser and Reflected Pulse Phase Delay 25

Beam Truncation and Collimating Lens Selection 29

Beam Shaping and Correction of Laser Diode Astigmatism 33

Optical Path Components 35

Focus and Tracking Servo Optics 37

Mechanical Design 39

ALIGNMENT OF THE OPTICAL HEAD 42

Optical and Mechanical Alignment 42

Focus and Tracking Detector Alignment 44

Alignment of the Beam to the Disk and Related Aberrations 46

RESULTS 50

CONCLUSION 55

APPENDIX A, Commercial Laser Diodes 56

APPENDIX B, Sharp LT024R10 Characteristics 59

APPENDIX C, WYKO Ladite Block Diagram 63

REFERENCES 65

5

LIST OF ILLUSTRATIONS

Figure 1. The Write Path and Coil 10

Figure 2. The Write Process 10

Figure 3. The Read Path and Polarization Diagrams1 12

Figure 4. The Complete Head and Read Path 14

Figure 5. Polarization Traces Showing the Effects of the Waveplates 15

Figure 6. Astigmatic Focus Servo Method1 18

Figure 7. Push Pull Radial Tracking1 19

Figure 8. Reflectivity Sensing WORM or CD Head1 21

Figure 9. Alternate Configuration of a Magneto-Optic Head7 24

Figure 10. High Frequency Modulation Timing Diagram11 27

Figure 11. Laser Noise versus Reflected Pulse Phase Delay10 28

Figure 12. Truncation versus Spot Diameter, Power and Irradiance 30

Figure 13. Astigmatism Introduced by Defocus in a 3X Circularizer 34

Figure 14. The Focus Error Signal 38

Figure 15 Threaded Mount and Barrel for Collimating Lens Positioning 40

Figure 16 Typical FES alignment sequence 45

Figure 17 Intensity Distribution of a Focused Spot Containing Coma16 47

Figure 18 Coma versus Tilt Angle Between the Beam and Disk Substrate 48

Figure 19 OPD Wavefront Plot and Beam Wavefront Error Values 51

Figure 20 Intensity Profiles of The Shaped Beam 52

Figure 21 Point Spread Function for the Shaped Beam 53

Figure 22 Photograph of the Optical Head 54

Figure 23 Laser Power versus Drive Current 60

Figure 24 Relative Intensity versus Theta Perpendicular 61

Figure 25 Relative Intensity versus Theta Parallel 62

6

LIST OF TABLES

Table 1. Oefocus, Axial Displacement, and Astigmatism 33

Table 2. 780nm Commercial Laser Diodes 57

Table 3. 830nm Commercial Laser Diodes 58

7

ABSTRACT

Design and operation of a modular optical head for a magneto-optic test system

are described. Alternate solutions to design problems are discussed. A 30mW semiconductor

laser with an integrated 250MHz oscillator is selected. The oscillator is used to modulate

laser read current for a reduction in laser feedback noise. A collimating lens with an

appropriate focal length is chosen so the beam's truncation at the objective yields the

maximum write power density. Astigmatism associated with the laser diode is reduced to

0.125 waves by defocusing the collimating lens and circularizing with an anamorphic prism

pair. Head components are aligned within several minutes of arc by using alignment

apertures and an autocollimator. Aberrations due to tilt between the disk and beam are

examined and coma is found to be the major contributor.

8

INTRODUCTION

Magneto-optic recording technology is the most successful erasable optical

scheme for computer disk data storage. The Optical Data Storage Center at the University

of Arizona's Optical Sciences Center carries out research on various aspects of magneto-

optic recording. The decision to build a magneto-optic disk test system for use at the

Optical Data Storage Center was made in March of 1988. The system's proposed capabilities

included media characterization as well as bit error testing. The system has since been

constructed and has demonstrated the abilities to read, write, and erase data. Hewlett-

Packard supplied the read/write electronics, test equipment, computer hardware, and system

software. The optics and mechanics for the magneto-optic test system were designed and

built at the Optical Sciences Center. This thesis is a description of the optical head, its

design, and fabrication.

Design of the optical head was influenced by several factors. The academic

research environment of the system made it necessary to design a modular head to simplify

future modifications. To allow for easy removal or replacement of the head, it was designed

with small components and constructed on a single platform.

Coordinate convention in this manuscript is as follows: the z-axis is the optical

axis, the x-axis is parallel to the plane containing the head components (horizontal), and the

y-axis is perpendicular to this plane (vertical). The following sections explain operating

principles of the head, various design tradeoffs, component selection, optimization, and

alignment procedures.

9

OPERATING PRINCIPLES OF THE MAGNETO-OPTIC HEAD

The Write Path

The writing process in magneto-optic recording is a thermal magnetic recording

process. During writing, the focused laser beam heats a small area (Dia. s; Iftm) of a

magnetic disk in the presence of an external magnetic bias field. The disk contains

magnetic moments perpendicular to its plane which align with the polarity of the external

bias field only if the temperature of the media is great enough. Figure 1 shows the write

path and the coil which produces the external bias field. Assume that all magnetic moments

on the disk are pointing in one direction as shown in Figure la. The coil's polarity produces

an external constant bias field (Hf) opposing the direction of the moments on the disk

(Figure 2b). The moments flip and align with the external bias field when the coercivity of

the media is lowered by laser heating. A magnetic domain (bit of information) is frozen

into the media when the laser's thermal energy is removed (Figure 2c), To write a stream of

data, this process is repeated by pulsing the laser rapidly while the disk spins. Typical laser

write powers at the disk are S to lOmW. Reversing the coil's field direction and heating the

media in a continuous path erases data. The moments along a track all point in the same

direction, and the initial state of Figure 2a is achieved. The media must pass under the

write head twice to rewrite over existing data, once to erase and a second time to rewrite.

The Read Path

The read process in magneto-optic recording is a polarization sensing technique.

The laser is on continuously and linearly polarized light is incident onto the magnetic media

as the disk spins. Upon reflection from the media, the polarization state is rotated an

10

Write Path

Beam Steering Mirrors

v

Collimating ^ Laser & Lens

1 Prism Pair

HF Modulator [ Taohs Actuator & Objective

Disk

Reference Detector

-Eg

•80:20 Partial Polarizing Beam Splitter

H, Coil

tan —* Hf

Figure 1. The Write Path and Coil

Written "Data' Laser Pulse,

Figure 2. The Write Process

11

amount 9* by the Kerr magneto-optic effect. Typical values of 6^ are 0.3 to 0.6 degrees.

The direction of rotation, clockwise or counterclockwise, depends on the magnetization

direction of the written domains in the media (Figure 2c).

A simplified read path is shown in Figure 3 along with corresponding

polarization diagrams. The dashed and solid arrows in the polarization diagrams depict

rotation from oppositely oriented domains (logical one's and zero's) on the disk. Note that

both states do not occur at the same instant in time. The 80:20 partial polarizing

beamsplitter has the following properties of transmittance and reflectance: Tp = 0.8, Rp =

0.2, Ts = 0, Rs «• 1.0. A description of the differential detection scheme is as follows.

Linear p-polarized light is incident upon the disk (a). Its polarization is rotated by ±8fc and

its amplitude is reduced after reflection (b). All of the s-component (the signal) and 20% of

the p-component is reflected by the partial polarizing beamsplitter (c). (A fraction of the p-

component is reflected because it is needed as a reference for the orientation of the s-

component. If Rp was large then Tp would be small, and the beamsplitter's power

throughput from the laser to the disk would significantly reduced write power.) If the

polarization shown in (c) was to go directly into the detection polarizing beamsplitter (A/2

plate removed) the differential amplifier would yield the same data signal for both a logical

one and a logical zero because the magnitudes of the p and ^-components are the same.

However, with the A/2 plate in place (oriented such that its fast axis makes an angle of 22.S

degrees with the horizontal), the plane of polarization is rotated 45 degrees (d) and the

differential amplifier's subtraction of the p-channel from the s-channel now produces a data

signal that is positive or negative depending on the direction of the Kerr rotation.

The partial polarizing beamsplitter used in combination with differential

detection produces an increase in signal by a factor of ISO when compared to the signal

generated by detecting only the .s-component of polarization rotation.1 This boost in signal

12

Simpl i f ied Read Path

Data Signal

Partial Polarizing Beam Splitter Tp =0.8, Ts=0

Rp=0.2, Rs = 1.0

0 * (a)'

(c) v/

13

is due to a first-order dependence on when using differential detection, as opposed to a

second order dependence on 0A when detecting only the s-component of polarization.

Another advantage of differential detection is common mode rejection, a characteristic of

the differential amplifier which suppress noise caused by intensity fluctuations.

The complete head is shown in Figure 4. The actual read path employs a A/4

plate and servo detection optics as shown. The A/4 plate improves the magneto-optic signal

by compensating for ellipticity in the polarization of the reflected beam. The following

explanation of how the A/4 plate effects the polarization is aided by the polarization traces

o f F i g u r e 5 [ A ] t h r o u g h [ F ] , a n d t h e c o r r e s p o n d i n g l a b e l e d p o i n t s i n F i g u r e 4 . F i g u r e s 5 [ A ] t

[5], and [C] show only the case of a +9* rotation representing a logical one. The initial

polarization state incident on the media is a linear p-component as previously shown in

Figure 3a. Figure 5 [A] shows the polarization components after the partial polarizing

beamsplitter and just before the A/4 plate. In reality |Er | » \Er^ | and Efp and Er^ are out

of phase by A^. Therefore, the resulting polarization is not linear, but slightly elliptical and

rotated from the p-axis by the small angle ±9^. The reflected beam then contains two

forms of information, the Kerr angle Qk and the amount of ellipticity (c, resulting from

The differential detection scheme as described above can only detect the Ken-

rotation, thus the information contained in the ellipticity is lost unless it is converted to a

useful form. Conversion of the ellipticity can be carried out by utilizing a Babinet-Soleil

compensator or by using a A/4 plate in conjunction with a A/2 plate. Both techniques yield

an equal increase in the signal to noise ratio. The Babinet-Soleil compensator retards the

phase of either the p or ^-component such that the two are in phase, producing a purely

linear state in which the 5-component (the signal) has been maximized. The method using

the A/4 plate deserves a more detailed description due to its complexity.

Light incident onto the A/4 plate has the polarization properties shown in

Figure 5 [A] and described above. The fast axis of the A/4 plate is oriented at 45 degrees to

14

Head Layout for M-0 Tester

Collimating Lens -i Laser Sc

HF Modulator

Taotis Actuator Sc Objective

Prism Pair Disk Beam

Steering Mirrors Reference

Coil

30:20 Partial Polarizing Beam Splitter

X/4 Plate [0.C.D1

A/2 Plate Quad. Oetectar

Astigmatic Lens Pair

Polarizing Beam Splitter

Data Channel 1 Sc Focus Error Signal Lens

Data Channel 2 Sc Tracking Error Signal

Quad. Detector

Figure 4. The Complete Head and Read Path

A. S

—E&-

+"e7 S.

t=0

t= l

t=2 t=3

Figure 5. Polarization Traces Showing the Effects of the Quarter and Half-Waveplates

16

the p and s-axes. Visualizing the p and ^-components as acted upon separately by the A/4

plate yields the right and left circular polarized traces as show in Figure 5 [£]. Vector

addition of the components which sweep out the opposite handed circles produces the ellipse

shown in Figure S [C]. Note that the major axis of this ellipse is at 45 degrees (ignore the

phase difference A for now) and the ellipticity ratio (the ratio of the minor axis to major

axis) depends only on the magnitude of the Kerr component, |£r |. Superimposed

polarization traces for +0* and -0A signals are shown in Figure 5 [£>]. After the A/4 plate

light passes through the A/2 plate (oriented such that its fast axis makes an angle of 22.5

degrees with the horizontal) which rotates the plane of polarization 45 degrees. For a

maximum difference signal the resulting traces should look like those shown in Figure 5 [F].

However, because of the initial phase difference A, the major axes of the ellipses do not lie

on the p and s-axes (see Figure 5 [£]). The A/2 plate is adjusted accordingly (away from its

22.5 degrees setting) to compensate for this phase difference by rotating the plane of

polarization so the axes of the ellipses lie on the p and s-axes to obtain a maximum signal to

noise ratio (see Figure 5 [F]). A mathematical description of this process as well as results

from increased signal to noise ratios can be found in reference 2 and 3,

Optical Servo Paths

Imperfections in disks and drive motors make it necessary to employ active

servo systems to keep the laser beam in focus and on track as the disk rotates. These servo

systems sense the position of the focused beam with respect to the disk and send electric

control signals to an actuator which supports and positions the objective lens. The actuator

responds by moving the objective in and out along the z-axis to maintain focus, and side to

side along the x-axis to keep the focused beam on track. The focus servo keeps the beam in

focus to ±0.5fim or less as it follows displacements of up to 500ftm. The tracking servo

17

maintains the beam's position on track to ±0.1 pm or better and can follow radial disk

runouts of 100/wi. The bandwidths of the tracking and focus servos are on the order of 1 to

5kHz.

The focus servo scheme used in the optical head was the astigmatic focus

method, shown in Figure 6. A spherical lens focuses the beam onto a quadrant detector and

a cylindrical lens introduces astigmatism into the beam. The quadrant detector is placed at

the circle of least confusion when the beam is in focus on the disk. All four quadrants of

the detector are uniformly illuminated and the focus error signal (A+C)-(B+D) is zero at the

in focus condition. When the disk moves in or out of the objective's focal plane the

reflected beam will converge or diverge. This effectively shifts the primary and secondary

image planes of the astigmatic system causing orthogonal ellipses to be imaged onto the

quadrant detector. The focus error signal generated, either positive or negative, is processed

by servo electronics and a control signal is sent to the actuator to focus the objective.

The tracking servo scheme employed was the push-pull radial tracking method,

illustrated in Figure 7. In this scheme diffracted light is reflected from grooves in the

surface of the disk. The grooves in the disk have a pitch (radial period) of 1.6/jm and an

optimum depth of A/8. The reflected beam from the media contains a bright zero-order

beam and two diffracted first-order beams which are equally spaced. Interference of the

first-order beams with the zero-order beam forms the baseball shaped interference pattern

shown in Figure 7. The resulting intensity patterns incident upon the quadrant detector are

shown in the figure. The tracking error signal generated is greater or less than zero

depending on the direction of the radial offset. As with the focus error signal, this signal is

processed and a control signal causes the actuator to correct the objective's radial position.

ASTIGMATIC FOCUS SERVO METHOD

Astigmatic Lens System Quad Detector Illumination

In focus

PUSH PULL RADIAL TRACKING METHOD

(A+O) - (B+C) >0 =0

20

DESIGN OF THE MAGNETO-OPTIC HEAD

Design Tradeoffs and Alternate Configurations

This section describes some of the reasoning that was done during the design

phase which led to the current head configuration. There are many forms a magneto-optic

head can take in response to several design problems. Advice on various issues was sought

from numerous experienced people in the field of magneto-optic recording. Their different

ideas and sometimes conflicting solutions to problems were appreciated. The following

paragraphs address several engineering tradeoffs and alternatives.

A mechanism was needed to translate the laser beam in x and y at the objective

as well as to control its angle with respect to the z-axis. It was decided that a simple v-

mount for the laser/beam-shaping assembly and two turning mirrors would do the job in a

compact fashion. One alterative was an x-y translation stage for the laser assembly and one

turning mirror for the beam angle adjustment. This would not have been as compact or

stable. The mirror mounts used were very stable flexure type mounts. These flexure

mounts are free from stick-slip problems, have low hysteresis, and are more stable over time

than the traditional ball bearing/spring-loaded thumb screw mounts.

The laser should be isolated from the reflected return beam to decrease laser

noise while reading data. The major source of laser noise is optical feedback into the laser

cavity from reflected light.4 This feedback causes instabilities and random competitive

mode-hopping in the cavity as well as a decreased power output.6*6 WORM and CD heads

employ a A/4 plate between the objective and polarizing beamsplitter to combat this problem

(see Figure 8). After passing through the A/4 plate, circular polarized light is incident on

the disk. Upon reflection the handedness of the light is changed and after a second pass

through the A/4 plate the polarization is rotated 90 degrees. The polarizing beamsplitter

optical disk

polarizing beam

splitter

circularizing optics collimating / Jens objective lens laser diode / lasi

(NT mirror p— astigmatic lens

quad, detector

Figure 8. Reflectivity Sensing WORM or CD Head1

N)

22

then reflects the return light towards the detection channel, isolating the laser from the

return beam. This will not work for a magneto-optic head since it is required that linear

polarized light be incident on the media.

Two options for controlling feedback induced laser noise were considered. In a

magneto-optic scheme, a polarizing filter and a Faraday rotator can be placed after the laser

and before the partial polarizing beamsplitter to yield optical isolation. Linear polarized

light from the laser passes through the polarizer and Faraday rotator, remaining linear while

its plane of polarization is rotated +45 degrees. The plane of polarization of the return

reflected light is rotated another +45 degrees after its second pass through the Faraday

rotator. This 90 degree rotation of the reflected light causes it to be absorbed by the

orthogonally oriented polarizing filter. This method was not used because of four problems

associated with Faraday rotators: cost, size, insertion loss, and isolation ratio.4

The second noise reduction technique examined was high frequency modulation

of the laser's DC drive current while reading data. A delay exist between the time a

semiconductor laser is turned on and the time it reaches single mode operation.6 High

frequency modulation (200MHz to 400MHz) of the laser's drive current from below

threshold to a level above threshold constrains the cavity to multimode operation. Though

the laser is not optically isolated, this method represses mode hopping caused by optical

feedback as well as mode hopping caused by thermal effects. The average optical power out

of the laser is much more constant and thus amplitude noise is reduced while reading data.

This technique is easy to implement (as discussed later) and has been shown to produce good

results.4

A common alternative to using a partial polarizing beamsplitter to reflect part

of the p-polarization towards the data channel, is to use a standard polarizing beamsplitter

rotated slightly in the plane of incidence. Rotation of a standard polarizing beamsplitter

causes the effect of a partial polarizing beamsplitter by reflecting a fraction of the p-

23

component towards the data channel. However, this technique complicates the alignment

and mechanical design of the head. Varying with the beamsplitter's degree of rotation are

two things: translation of the beam in the aperture of the objective, and the angle of the

reflected beam towards the data channel. The use of a partial polarizing beamsplitter made

the mechanical design straightforward and simple.

The head of Figure 9 illustrates three additional aspects which were considered

during the design period but were rejected. The first aspect shown is the disk mounted

horizontal, parallel to the plane of the head. Horizontal mounting of the disk is common in

most disk test systems. Vertical mounting was decided upon for two reasons. If a glass disk

shatters in a vertical plane it is less likely to hit and injure someone than if it shatters in a

horizontal plane. Also, the head optics are more accessible and easily viewed with a vertical

configuration.

The second aspect was the separation of the servo channel from the data

channel via an additional beamsplitter. Data and servo channels are sometimes separated to

eliminate feedthrough from the servo signals into the data signal. Feedthrough was not

estimated to be a problem with the servo techniques chosen. The decision to combine the

data channel and the servo channel simplified the head by reducing the number of optical

and electrical components. Another benefit gained was a shorter write path which decreases

laser noise as described later in the discussion of reflected pulse phase delay.

The third aspect illustrated in Figure 9 is the 45 degrees rotation of the data

channel instead of using a A/2 plate to rotate the plane of polarization. Rotation of the data

channel was decided against for mechanical reasons. It is much easier to rotate and adjust a

single A/2 plate than to mount the data channel on a separate rotatable mount and baseplate.

It is also easier to keep the data channel aligned if it is mounted in the same plane as the

other head components. The A/2 plate thus led to a simple, more stable and condensed

design.

(MO-)Konol

Figure 9. Alternate Configuration of a Magneto-Optic Head7

to

25

Another design alternative was to use a Babinet-Soleil compensator instead of

the A/4 plate. This was decided against because of the size and cost of a Babinet-Soleil

compensator.

The choice of astigmatic focus and push pull tracking for the servo systems was

made because these techniques have been researched extensively and are widely used.1*8

They are also fairly simple in principle and implementation.

The Semiconductor Laser and Reflected Pulse Phase Delay

Magneto-optic recording requires the laser diode to have an output power of

20mW to 30mW so that about lOmW is incident onto the media during the write process. It

was decided that the laser diode for the magneto-optic head would be index-guided and

emit at 780nm. Index-guided lasers tend to have better beam properties, sharper turn on,

and more linear output/current characteristics than gain-guided lasers.9 Astigmatism

associated with index-guided lasers is usually 3pm to 5fim, whereas astigmatism associated

with gain-guided lasers is typically 20pm to 60pm. Also, the intensity profile of gain-

guided lasers is less Gaussian than that of index-guided lasers and index-guided lasers are

less prone to feedback.4'9 Index-guided lasers are more desirable for applications that

require a small spot size.

A 780nm laser diode was preferred over an 830nm laser diode because reliable

high power 780nm laser diodes are now available, and areal storage capacity increases as the

square of the wavelength. Lists of characteristics and prices for commercially available high

power laser diodes emitting at 780nm and 830nm can be found in Appendix A.

The laser chosen for the head was an index-guided, 780nm Sharp LT024R10

with a maximum output power of 30mW. Specifications and characterization data for this

laser are included in Appendix B. This laser is packaged with a 250MHz integrated

26

oscillator which modulates the read current (rDCread) down to a hard turn off below

threshold. This is illustrated in Figure 10; the optical power out (P0) is modulated between

zero and some read level when the modulator control signal (Vcc) is turned on. This high

frequency modulation forces the diode to operate multimode, reducing intensity fluctuations

caused by mode hopping, as previously described.

The relation between read current modulation frequency and the laser-to-disk

optical path length, has a strong effect on laser feedback noise. (Do not confuse the

modulation frequency here with modulation of the laser during writing. This is solely a

read current modulation technique used to reduce noise by modulating the normally constant

laser read current.) More specifically, laser feedback noise is a strong function of the phase

difference between the modulated irradiance leaving the laser and the reflected irradiance

returning to the laser.10 This noise is smallest if the reflected intensity at the laser is

minimum when the laser's emitted power reaches a maximum. If however, the reflected

intensity at the laser is maximum just before the laser's emitted power reaches a maximum,

the laser is unstable and the feedback induced noise is large. An expression describing the

phase delay of the reflected laser read pulse is given by:10

RPPD = /^K360 degrees) [1]

where RPPD is the Reflected Pulse Phase Delay, / is the modulation frequency, L is the

optical path length from the laser cavity to the reflective disk, and c is the speed of light in

air. (The phase delay described here is based on the electrical modulation frequency; it is

not an interferometric phase delay based on the wavelength of light.)

The frequency measured for the modulator of the Sharp LT024R10 was / =

249.4MHz. The optical path length was minimized to yield L = 8 inches. Evaluating the

expression above with these values yielded a reflected pulse phase delay of 122 degrees.

Figure 11 is a plot of the laser noise to shot noise ratio versus the reflected pulse phase

delay. The circles, triangles, and squares represent modulation frequencies of 200, 300, and

nruu 20mW

J u HflAA/l/l/yyVWi/lA/IA/M u L

High frequency modulator ON

Figure 10. High Frequency Modulation Timing Diagram11

10000

K\ H QV

A nu

180 360 540 720 000

phase delay (degrees)

Figure 11. Laser Noise versus Reflected Pulse Phase Delay10

29

400MHz respectively. At 122 degrees the laser noise to shot noise ratio is below 10 on the

logarithmic scale, a respectable value. Note that the minimums in noise repeat every 360

degrees - at 90 degrees, 450 degrees, etc. To obtain a reflected pulse phase delay of 90

degrees, with the fixed frequency of the Sharp modulator, the optical path length required is

about 6 inches. This was too constraining for the small amount of noise reduction gained.

For a reflected pulse phase delay of 450 degrees the path length would need to be 30 inches.

Varying the frequency of the modulator allows easy achievement of a 90 degree

reflected pulse phase delay and relieves the requirement of optimizing the optical path

length. This requires an external variable high frequency oscillator to modulate the laser's

drive current during the read operation. One problem with this is that the cables from the

high frequency oscillator to the laser radiate interference that shows up as noise in the data

detection electronics.13 This problem is minimized with the Sharp LT024R10, because the

oscillator and laser are packaged in a common shielded case and the lead lengths are minute.

The Sharp LT024R10 was an inexpensive, easily implemented off-the-shelf solution to read

current modulation.

Beam Truncation and Collimating Lens Selection

Two factors guided the selection of a lens to collimate the laser's divergent

output the size of the focusing objective's aperture, and the wider divergence angle of the

laser's elliptical output beam (Qperp). Filling the objective lens aperture involves tradeoffs

illustrated by Figure 12. Underfilling the objective considerably with the Gaussian beam's

1/e1 profile effectively reduces the NAobj. This produces a larger spot size and lower power

density on the disk. However, transmitted power is very high. Conversely, if the aperture

is considerably overfilled (uniformly illuminated), the smallest possible spot size is obtained.

In this case the power density and transmitted power are low. At a point between underfill

I '

Irradiance

Transmitted Power

CO 0.5

0 0.5 1.0 1.5 2.0

Truncation: DIA,en9 -5- i/e2 DIAgau88jan

Figure 12. Truncation versus Spot Diameter, Transmitted Power and Peak Irradiance1 o

31

and overfill the power density reaches a maximum. This occurs when the aperture is

slightly underfilled at the truncation ratio given by:1

_DIA \ e*

objective g 1,12 [2]

• ^^Gaussian

As shown below, this ratio is equivalent to truncating the Gaussian at the 1/e*-6 intensity

level. It was decided that the objective's aperture would be filled to this extent because of

the high power density obtained and small sacrifice in transmitted power and spot size.

The objective's aperture was first matched to the beam's intensity

characteristics. The focal length of the collimating lens to make this match was then

calculated and a collimating lens was selected. This process is outlined as follows. A

radially symmetric Gaussian beam intensity profile is given by:

/(r) » I0e

32

R - l°»i TM [ioj

The necessary focal length of the collimating lens (fcou) can then be calculated from

OT '»«" r I")

sinfWj

/coll = U, NA°"I , [12]

Substituting R from equation [10] into the above equation yields

NAnM

1.90 sinjW-j

The objective in the Taohs-PB7 actuator has a focal length of fobj = 4.3mm and a NA of

NAobj = 0.5. The Sharp LT024R10 laser's 9perp is 29 degrees. Evaluating expression [12]

with these values, the desired focal length of the collimating lens was fcoH » 4.52mm. The

only other criterion to meet in selecting the collimating lens was that its aperture be larger

than that of the objective's so truncation by the collimator would riot occur. The objective's

aperture was 4.3mm. The collimating lens selected has a focal length of 4.52mm, an

aperture diameter of 4.3mm, and an NAcon of 0.48. Inspection of the curves in Figure 12

about the dashed optimization line reveals that exact optimization is not critical due to the

flatness of the curves in this region.

It was not surprising that the collimating lens' focal length and diameter were

precisely what was needed since the Taohs actuator/objective is one of the most popular

available. The collimating lens, distributed by Optima, is a four element air spaced lens

with wavefront aberration typically less than A/10. It was designed for 785nm and all

surfaces are magnesium fluoride anti-reflection coated. The price is amazingly low at

$25.00 each in small quantities.

33

Beam Shaping and Correction of Laser Diode Astigmatism

The circularizer chosen was a Melles Griot anamorphic prism pair with a fixed

ellipticity correction ratio of 3 to 1. This was close to the laser's specified ellipticity ratio of

2.9 to 1. The circularizer stretches the narrow axis of the elliptical beam after collimation,

giving rise to a circular beam. If the input to the circularizer is an elliptically truncated

plane wavefront, the output will also be a plane wavefront, since a plane stretched by three

time in one direction is still a plane. However, if the input is a spherical or astigmatic

wavefront, the circularizer can introduce or cancel astigmatism by stretching one axis of the

wavefront. The following is an explanation of how the laser's inherent astigmatism can be

corrected by defocusing the collimating lens and allowing the beam to pass through the

anamorphic prism pair.

The mechanical requirements of positioning the focal point of the collimating

lens in proximity to the laser diode's facet are stringent. The relation between defocus (lf20)

in waves and the axial distance Az is given by the relationship,

= [13]

For A » 780nm and an NAcotl = 0.48 the values below are obtained.

( * ) A t ( n m ) W „ ( A )

1 6 . 9 0 . 8 9 0 . 1 0 . 6 9 . 0 8 9 0 . 1 4 1 . 0 . 1 2

Table 1. Defocus, Axial Displacement, and Astigmatism

Ignore the inherent astigmatism in the laser diode for now. Astigmatism (Wj2) induced by

the 3X circularizer is related to defocus by W22 = (8/9)W20 as illustrated in Figure 13.

Defocus produces a spherical wavefront and is not as harmful to the system as astigmatism

since the focus servo will keep the scanning spot in focus despite a slight divergence in the

beam. From the table above it is realized that in order to keep astigmatism to about A/10

34

Astigmatism = (8/9) Defocus For a 3X Prism Circularizer

OPD Contours - Defocus No Astigmatism

Shape of Laser Diode's Elliptical Beam Truncates Wavefront

OPD Cross Sections Truncated by Ellipse

vY

I V,x 7 " ' T

B

H d/3 |«- h

A=B(1/9)

After 3X Circularizer the Beam Is streched 3X along X A and B are still the Same

Sag = Astigmatism = B-A = B(8/9)

*

B is due to Defocus So Astigmatism=(8/9)Defocus

Figaro 13. Aatigmatism Introduced by Defocus in a 3X Circularizer

35

the collimating lens must be positioned within about 1 fim of its optimum location. A

mechanism which allows for this degree of control is described later in the section on

mechanical design. Owing to the relation between defocus and astigmatism as illustrated in

Figure 13, it was possible to correct for the laser's inherent astigmatism (an astigmatic length

of £! 3fim) by adjusting the focus of the collimating lens. As the collimating lens was

defocused the resulting astigmatism was minimized by monitoring its magnitude on a

WYKO Ladite phase shifting interferometer (see block diagram in Appendix C). The results

are presented later.

Optical Path Components

Several optical path components which deserve a brief treatment are: the

magnetic bias coil, the reference detector, the Taohs actuator, the 80:20 partial polarizing

beamsplitter, the waveplates, and the detection polarizing beamsplitter.

A magnetic bias coil can be made by winding magnet wire around a piece of

"soft" steel. An alternative is to use the coil from a 12 volt relay. A relay coil is precisely

wound, very neat and cheap. A coil from a relay was used in the test head. The coil was

calibrated with a Gauss meter to measure the magnetic field a fixed distance away as the

current through the coil was varied. Typical field strength at the disk is 300 Gauss.

The reference detector is a simple silicon detector just large enough to contain

the beam without focusing optics. It was tilted to avoid reflection of light back to the laser.

This external detector can be used to monitor the laser power in a feedback loop to stabilize

the laser's output power. The photodiode built into the laser diode is not used because it can

give false power readings due to the difference in emission between the front and back

facets. Temperature variations of the laser package and noise in the laser caused by

reflections can also effect the internal photodiode's accuracy.

36

The Olympus Taohs actuator was chosen mainly because of its proven reliability

over the years.13 Its size is large compared to other actuators but it is still used in WORM

drives today since it is dependable. Another plus was that the electrical engineering support

at Hewlett-Packard was familiar with its electrical drive requirements.

The 80:20 partial polarizing beamsplitter was donated by OCLI. It has

polarization phase control built into its coatings which are non-symmetric. The orientation

of the cube must be correct for it to act as desired with minimal phase retardation of the

polarization. The exact ratio of Tp to Rp is not critical but it should be realized that the

fraction, Rp, of the power incident on this beamsplitter is lost on the way to the disk. This

beamsplitter also cleans up the laser's polarization before the disk, reflecting any unwanted

s-polarization.

The half and quarter-waveplates selected were zero-order optically contacted

waveplates. Zero-order waveplates are less sensitive to wavelength variations, making them

ideal for laser diode applications. A ±2% deviation from the center wavelength results in a

0.01 A retardance error (3.6 degrees). These waveplates are also fairly insensitive to

temperature changes. A typical temperature coefficient is 0.0001A per degree Celsius. The

waveplates, purchased from Karl Lambrecht, are AR coated and mounted in a small barrel.

The detection polarizing beamsplitter is a half inch cube, also from Karl

Lambrecht. Its extinction ratio is 500 or better, and it is AR coated as well. The read path

function of the servo optics is to focus light onto the quadrant detectors. The signals from

the quadrants of each detector are summed for the purpose of the differential data channel.

If the servo channel was separated from the data channel, the data channel detectors would

be single element detectors.

37

Focus and Tracking Servo Optics

The servo optics consist of an astigmatic lens pair in the focusing channel and a

single element lens in the tracking channel. The servo function of the astigmatic lens pair is

to introduce astigmatism for focus error detection. These elements are AR coated one inch

optics, purchased from CVI. The spherical lens of the astigmatic lens pair and the single

element lens in the tracking channel both have a focal length of 100mm. The cylindrical

lens has a focal length of 150mm. The active area of the quadrant detectors is 1mm1; each

quadrant is 0.5mm on a side. The gap between quadrants is 10/wi. These detector were

purchased from United Detector Technology.

The focus servo design was done by Scott DeVore with software he had

previously developed.14 The spherical lens is placed before the cylindrical lens in the beam

path. The thin lens separation is 46.75mmt and the distance from the rear principal plane of

the cylindrical lens to the detector plane is 45.22mm. The characteristics expected from the

design model are shown in Figure 14. Figure 14a is the focus error signal, (A+C)-(B+D)

normalized by the sum. The horizontal axis of this plot is the displacement of the disk.

Figure 14b shows the anticipated spot on the detector at five equally spaced increments

centered at the zero crossing of the focus error signal. The focus error signal is not

symmetric about the zero crossing, thus the blurs are not symmetric.

The tracking servo was not simulated since the diffraction pattern is entirely

dependent upon the various groove structures of the different media used. The only

requirement is that the spot is centered on the detector. A spot diameter roughly equal to

the "in focus" spot on the focus servo detector is desired. A 100mm lens was used to project

the spot onto the detector since the same focal length lens was used in the focus servo

channel. This was done to keep the two servo channels as symmetric as possible for the sake

of differential detection.

38

1.00

•f§t ilHUUflMi'ftuii ErMA ilfiWL " ' 117577!?-Univarsllv of AZ Itodla T«*t«r — ISO m cvl 17:2C:1S

/\^

-1.00 -49.00 FOCU ERROR (unl 43.00

res 1IHULAT1W: GETECTOA BUJA vm KPUUt Unlwralty of AZ Had la Tutor — ISO m cvl

il/07/lf 17:30:31

A / YW ' X? ; y ' /

B

Figure 14. The Focus Error Signal (a), and The Intensity on the Detector at Focus and at Two Equally Spaced Locations On Each Side (b)

39

Mechanical Design

The mechanical design was the most time consuming aspect of the project.

Since modularity was a key issue, the majority of components were mounted separately

within the limits of practicality. The mounts that were designed and machined accounted

for 65% of the component mounts. Of the mounts that were purchased, 50% required

machining for modifications. Some of the more interesting aspects of the mechanical design

are discussed below.

Micron control of the distance between the collimating lens and the laser's facet

was needed to correct for astigmatism using defocus, as previously mentioned. Two

mechanical parts that provided this control are the mount and threaded barrel of Figure 15.

The threaded barrel that holds the collimating lens is brass and the threaded mount is

aluminum. The threads on the O.D. of the barrel and I.D. of the mount are 80 pitch, class

3A threads. Three angular degrees of rotation is a small rotation to achieve with a steady

hand and a screwdriver, but one degree is difficult without a special lever arm. A three

degree rotation of an 80 pitch thread yields a 2.6/un displacement and a one degree rotation

yields a 0.9tim displacement. A special lever arm was designed and it was decided that

iterative trial and error positioning of the collimating lens would be necessary. Thermal

calculations for a 5mm length of aluminum revealed a displacement of O.lnm/K, an

acceptable value over the expected changes in room temperature. Once in position, the

threaded barrel was held in place by glue. A low viscosity, low shrinkage glue was injected

- into the glue hole. A set screw was not appropriate because tightening the setscrew would

displace the barrel more than a micron. The threaded mount fits into a circular hole in the

barrel of the anamorphic prism beam shaper and is held in place by a set screw. The laser

is attached to the threaded mount via two small screws.

Most of the mounts were machined out of standard 6061-T6 rolled aluminum.

40

Mount

Laser Diocle

Threaded Barrel

-Glue Hole

Bean Shaper

•CoUlnating Lens

Figure 15 Threaded Mount and Barrel for Positioning of the Collimating Lens

41

The baseplate for the head was made of tooling plate aluminum. This is a cast aluminum (as

opposed to rolled) which is stiffer and has better thermal properties than 6061-T6

aluminum.

The baseplate of the head is attached to the head translation stages through a

semi-kinematic mount. This mount allows removal and remounting of the head without

realignment. The angular repeatable accuracy of the head with respect to the disk was

measured to be 15 seconds of arc.

Many of the mechanical parts were designed on a computer with a mechanical

CAD drawing package. The overlay or layer capabilities of the CAD package proved to be

invaluable, especially for the more complex parts and assemblies. Design error was reduced

through a visual verification of the proper mating of parts.

42

ALIGNMENT OF THE OPTICAL HEAD

Optical and Mechanical Alignment

The alignment of the laser and beam shaping assembly was done at WYKO

Corporation with the assistance of the Ladite interferometer. The remaining components

were aligned as they were fastened to the head's baseplate with the aid of an autocollimator

and an alignment aperture. The detectors and actuator were mounted on small xyz-mounts

which were adjusted during operation of the head while viewing signals on oscilloscopes.

The laser was centered in the aperture of the collimating lens by measuring the

beam position at the collimating lens, and then about eight feet away in the +z direction.

Positioning it was not as difficult as trying to compensate for its movement when the two

mounting screws were tightened. Once the laser was centered, the beam was roughly

collimated by inspection of the spot size near and far from the laser. Alignment and

collimation was done at the laser's read power since astigmatism changes with output power

and it is during reading that we are most concerned with beam quality. The following

iterative procedure is outlined with numbered steps.

1) The laser and collimating lens mount were attached to the circularizer and mounted to

the Ladite.

2) The laser and collimating lens mount assembly were rotated with respect to the

circularizer while viewing the intensity pattern in real time. When the intensity pattern was

circular the circularized setscrew fixed the mount's position.

3) Data was taken on the amount of astigmatism and the RMS wavefront error with tilt and

defocus removed.

4) The laser and collimating lens mount were then removed from the circularizer and the

collimating lens barrel was defocused slightly by tapping the threaded barrel to produce a

43

small rotation.

5) Go to step one and repeat procedure.

This was done for about 12 iterations until a low value of astigmatism was obtained through

the proper amount of defocus (see results section). The collimating lens position was fixed

by gluing its threaded barrel in place. Tilt and defocus were removed from calculation of

the RMS wavefront error since tilt depends on the alignment of the head with respect to the

disk, and defocus is corrected by the focus servo and actuator.

Rough alignment of the beam on the optical head with respect to various

components was accomplished by passing the beam through an alignment aperture and

watching the reflected beam on the periphery of the aperture. When the beam passed back

through aperture the reflective component was roughly in the right place. The alignment

aperture was a quarter inch diameter rod with a 0.05 inch diameter hole at the optical axis

height. Two alignment apertures were made. The baseplate was designed with mounting

holes for these apertures at proper locations.

The alignment apertures were used to get a component close enough to its

desired position so that a reflection off of the component could be seen in the field of view

of the autocollimator. With the autocollimator, the component could be positioned within

several minutes of arc. Some of the components were shimmed with brass shim stock for

fine angular adjustments. The following list gives the positional accuracy of several head

components. The autocollimator's vertical and horizontal cross hair's angular displacements

are denoted with V and H. WRC implies that the alignment was Within the Resolution of

the Cross hairs (about IS seconds of arc or less). Minutes of arc is denoted by "min".

1) P-polarization out of the laser was ±2 degrees to the horizontal

2) 80:20 PBS - Hi 3 min, V: WRC

3) Actuator mounting flange, over x^-range of travel - H: WRC, V: 3 min

4) A/4 plate - ff: WRC, V: WRC

44

5) A/2 plate - H: 3 min, V: 3 min

6) Detector PBS - H: 0.5 rain, V: 0.5 min

7) Disk w.r.t. Beam H: WRC, V: 1 min

Focus and Tracking Detector Alignment

The quadrant detectors were initially aligned in x and y by eye. The focus

servo was aligned by slightly tilting a reflective non-grooved disk and spinning it so it

moved inside and outside of focus. The focus actuator could also have been ramped in and

out of focus with a function generator. With the disk wobbling the actuator was disabled.

All four quadrants of the focus detector were connected to a four channel scope. The

detector's position was varied until four responses of approximately equal magnitude were

seen on the scope. The detector was moved around while trying to get the traces of diagonal

quadrants, (A&C, B&D) to overlap (the cylindrical lens was at 45 degrees to the horizontal).

Overlapping of diagonal quadrants signified that these quadrants were equally illuminated

and that the spot was centered. Quadrant scope traces looked like those shown in Figure 16.

If the diagonal quadrant's traces crossed each other, this signified that the spot was being

translated on the detector as the disk went through focus. This occurs when the beam is not

centered in the objective; the chief ray is displaced from the center of the objective. The

objective was then translated to uncross the traces. This moved the spot on the detector and

the detector was repositioned until the diagonal quadrants' traces overlapped. This method

was iterated until the diagonal's traces overlapped and did not cross. The focus servo was

then able to lock on focus. A detailed and well illustrated method for adjusting an

astigmatic focus servo can be found in reference 15.

To align the tracking servo a reflective non- grooved disk was placed on the

spindle. Each quadrant of the tracking detector was connected to a separate channel of a

Figure 16 Typical alignment sequence using a quadrant detector. Both FES and quadrant signals are shown. The alignment progresses from a centered detector with a moving blur (a and b) to a stationary blur that is no longer centered (c and d). At this point the detector needs only to be moved to the center of the blur, and the alignment will be complete.15

46

four channel oscilloscope. The detector was centered by equalizing the response of all

quadrants when the beam fell on the detector. A grooved disk was placed on the spindle

and the tracking servo locked up after some adjustments in the electronics.

Alignment of the Beam to the Disk and Related Aberrations

It is critical that the scanning beam be normal to the disk. A tilted disk or

misaligned head can cause read back errors due to aberrations introduced by the disk's

protective layer and clear substrate. These layers, treated as one substrate, act as a plane

parallel plate with a typical refractive index of n = 1.57 and a thickness of T = 1.2mm.18

Coma can be introduced by a decentered or tilted objective lens, but it is much more

sensitive to the angle of tilt between the focused beam and the disk substrate. Astigmatism

is also introduced from this tilt but it is negligible compared to the amount of coma

introduced by the same tilt angle. The expressions for coma (1K31) and astigmatism (Wiz) in

terms of waves are given by:16*™

= t14l

n^r(NA)2^ ll5]

where yS is the tilt angle in radians and T is the substrate thickness. If the coma lobe (see

Figure 17) of the spot on the disk is oriented towards neighboring tracks it can cause cross

talk16 as well as tracking errors." A quarter wave of coma is about the limit that the system

can be tolerate.17 A plot of Wai verses /9 in degrees, is shown in Figure 18. Values from

the head, used in the plot were A=780nm, «=1.57, 7=1.2mm, and NAobj=0.5. The plot shows

that a tilt of 0.4 degrees produces a quarter wave of coma. The amount of astigmatism

introduced at this same angle is -0.003 waves, two orders of magnitude less.

Spherical aberration introduced by the plane parallel plate substrate is given

47

Figure 17 Intensity Distribution of a Focused Spot Containing Coma18

48

yd

7

0.4 0.6 0.8 1.0 1.2 Beta (degrees)

1.4 1.6

Figure 18 Coma versus Tilt Angle Between the Beam and Disk Substrate

49

by:10

- T̂(NA>'

The objective lens is designed to correct for this by introducing the same amount of

spherical aberration of the opposite sign. The amount of spherical aberration present

without correction is very large: 4.5 waves for A=780nm, n=1.57, r=1.2mm, and NAobj=0.5.

Although the objective was designed to correct for this, allowed thickness variations in the

substrate of ±0.05mm18 can produce ± 0.2 waves of spherical aberration. The objective lens

in the Taohs actuator was designed for «=1.5, not m=1.57. This difference combined with

the thickness variations above yields a worst case of 0.29 waves of spherical aberration

introduced. This would reduce the peak intensity of the focused spot a considerable amount

(roughly 20%) if no refocusing was done.

The astigmatic focus servo works to reduce this aberration by introducing

approximately -0.65 waves of defocus for every one wave of spherical aberration.16 The

peak intensity after refocusing is only reduced by about 5%.1B The spot quality is then

considerably better than if no refocusing were done. With spherical aberration introduced in

the reflected beam, the focus servo tends to refocusing more towards the marginal side of

focus than the paraxial side.

50

RESULTS

An OPD plot and wavefront data from the WYKO Ladite are shown in Figure

19. The astigmatism value of 0.125 waves in the lower half of the figure was the value that

was monitored and minimized while adjusting the defocus. The RMS wavefront error is

0.029 as measured out of the circularizer. If the RMS wavefront error before the objective

is less than 0.05 it does not have a significant effect on the spot size.19 The intensity

profiles and point spread function are shown in Figures 20 and 21. The horizontal axis of

the intensity profiles is scaled such that 1 => 3mm. The vertical axis is not normalized to

unity at the peak intensity and thus the 1/e2 points are not as indicated. The \/el diameter

is larger than that shown in the figure.

The assembled head is shown in Figure 22. The actuator and objective

assembly is the round object just above and to the right of the watch band. Above and to

the right of this, with wires coming off of it, is the laser diode and built-in modulator. On

the left of the figure three electronics boards can be seen. The larger horizontal board to

the left is a data channel board. The two smaller boards are preamplifier boards for the

quadrant detectors.

At the time of this writing the tester has not been calibrated and thus detailed

test results are not available. The optical head has been shown to read, write, and erase data

with a carrier to noise ratio of 50dB. This carrier to noise ratio was obtained without the

A/4 plate in place and without the laser read current modulated. Both the A/4 plate and the

modulation of the read current will increase the carrier to noise ratio. The current

configuration of the laser driver does not allow operation of the modulator.

Neui Data

r ois : 0 . 02 9 u / v : 7 0 0 . 0 n m

0 3 : 4 2 0 7 / 0 9 / 3 9

OPD

-a.eg -0.23

T F

p — v : 0 . 3 1 3 pupil: 3 7'<

O r l « n t * t l o n

• r

C T r o n t 1

a.a?

9, SI

a. i?0

o. 093

- a . BBS

-ft.144 Ul 1, tr-*0 Li -O.144

WYKO

LADITE CVer 4.21] Today's Data: 07/09/89

New Data Tina: 63:42 Oats: 67/09/89

SE]DEL ABERRATION COEFFICIENTS OF KWEFRONT: AMT ANGLE TJ LT

FOCUS AST1C com SPHERICAL

2.275 0.177 0.125 0.077 0.023

-259.8

-38.7 -47.9

LONGITUOINAL ASTIGMATISM 0.9 ua

ABERRATIONS REMOVED: TILT FOCUS

X CENTER V CENTER

129.0 108.5

RADIUS

90.8

PUPIL: 97X

OATfl PTS PEAK

23144 0.176

VALLEY P-V -6.144 0.313

RHS 0.029

ftPOOJZED UEOGE STREHL RATIO 1.00 0.979

WYKO

Figure 19 OPD Wavefront Plot and Wavefront Error Values of The Collimated and Circularized Laser Diode Beam

New Da ta 03 :42 07 /09 /89

intensity ;:;,m0,s;s f l p t r t u r > H t d t h : t . A S X P R O r i L E 0 . 7 6 0

0 . 3 B 0 -

0. 000 1 . 0 -0.5 0.0 0 .5 1 . 0

R e l a t i v e D i s t a n c e i n P u p i l

MYKO

Neui Da ta 03 :42 07 /09 /09

INTENSITY p^r-9™ B p i r t u r a H t d t h t 1 . H f l Y P R O F I L 0 . 3 7 4

0 . 2 1 9 -

0. 000 0 .5 1 .0 0 .0 1 . 0 -0 .5

R e l a t i v e D i s t a n c e i n P u p i l

MYKO

Figure 20 Intensity Profiles of The Shaped Beam Note; Vertical Scale is Not Normalized

N e t u D a t a

S t r e h 1 : 0 . 9 7 3 5 i : s ! 9 . 3 6 u m f l p o d i z a d

0*J: 42 07/09/89

PSF

l /•

T F

u j » / : ? 9 0 . O n m pupil: 3 7

p l o t I C l I t ! Q . 9 7 Q t t r a h l u n i t s

Q d e g 9 0 d e g

- 4 5 d e g

WYKO

N e w D a t a

3s 0) L © C a TJ

54

55

CONCLUSION

Operating principles of the head were described in terms of the write path, read

path, and servo path. Design alternatives were then reviewed and design decisions were

justified. Key issues of the optical head design were presented. Among these were an

explanation of reflected pulse phase delay, collimating lens selection for optimum beam

truncation at the objective, and correction of laser diode astigmatism. Alignment issues

were then addressed.

The optical head's capabilities to read, write, and erase data were successfully

demonstrated. The carrier to noise ratio obtained without complete optimization of circuitry

and optics is very promising. Meaningful test results will follow the complete calibration

and interfacing of the system's optics, electronics, and software.

APPENDIX A

Commercial Laser Diodes

HIGH POWER LASER DIODES 780 nm

Mfg. G=Gain Max 0*@ x Astig Price Pt.# I=Index P™ 6° 0£ Pout A (urn) (U.S.Dol)

Guided (raW) (mW) fom) 1 pc.

HITACHI 7838G I 20 10 26 20 780 146.00

MITSUBISHI ML6702A ! 30 12 23 30 780 175.00

6412N/C I 30 12 21 30 780 175.00 6411A/C 20 12 30 10 780 81.70

TOSHIBA Told

121/122 I 20 10 27 20 780 package 127 I 20 10 27 20 780 121T I 30 9 20 30 780

SHARP LT024 I 30 10 29 20 780 3 260.00 LT024R10 I HF modulation built in 389.00

SONY SLD201U-3 G 50 14 28 40 780 10 180.00 SLD201V-3 G 50 14 28 40 780 40 180.00 SLD201U G 20 13 28 15 780 4-10 SLD20IV G 20 13 28 15 780 40-60

Table 2 780nm Commercial Laser Diodes

HIGH POWER LASER DIODES - 830 nm

Mfg. & Mndex Max fif@ Pries Part# G-Gain P„ 9® Si P« A Astig (U.S.Dol)

Guided (mW) (mW) (nm) (/ttm) 1 pc.

NEC NDL3I10 30 12 30 830 2 S

hitachi 8351E 50 9 27 50 830 8.5 1,030.00 8318E 40 11 25 40 830 4 560.00 8312E 20 10 27 10 830 256.00 8312G 20 10 27 10 830 256.00 83143E/G 30 10 27 30 830 307.00 83J5E 20 10 27 10 830 1,030.00

qrtel XL303 30 30 30 30 835 1,150.00

MITSUBISHI ML5702A I 30 12 21 30 820 175.00

5412N I 30 12 23 30 820 175.00 5415N I 30 11 26 30 825 407.50

toshiba Told

131/132 I 35 10 27 35 830 131TR I 40 9 20 40 830 131T I31TV I 35 9 20 35 830 13SA I 30 10 27 25 830 136 A I 25 10 27 20 830 137 I 40 10 27 35 830 151/152 I 35 10 27 30 810 153/154 I 25 10 27 20 810 155 A I 25 10 27 20 810 157 I 35 10 27 30 810

sharp LT015 I 40 10 27 30 830 LTO15RI0 I 40 10 27 30 830 LT016 I 40 810

PHILIPS amperex CQL61A G 20 35 40 20 820 20 CQL62A G 40 30 38 40 820 20

SONY SLD202U-3 G 50 13 28 40 820 SLD202V-3 G 50 13 28 40 820 7 SLD20V G 25 13 28 20 820 SO

Table 3 830nm Commercial Laser Diodes

APPENDIX B

Sharp LT024R10 Characteristics

60

Optical Power Out VS. Input Current Sharp LT024R10 Laser # 3

o. ro

if).

61

Sharp LT024R10 Laser Diode No. 3

1.00 -I

oo

— 0.50 -Ld

o.oo TTT ' ' • I | I -40.0-30.0-20.0-10.0 0.0 10.0 20.0 30.0 40.0

THETA PERPENDICULAR

Figure 24 Relative latensity versus Theta Perpendicular

62

Sharp LT024R10 Laser Diode No. 3

1.00 H

cn

L±J

^ 0.50 -LLI

h ;

0.00 -10.0 -5.0 0.0 5.0 10.0

THETA PARALLEL

Figure 25 Relative Intensity versus Theta Parallel

APPENDIX C

WYKO Ladite Block Diagram

DETECTOR

FOUNNQ ynttvt «T FOCUSN6 LOIS

BEAM SPUTTER ft

FUP-IM BEAU BLOCK| SHUTTER k—if

4?=Q= REFERENCE ARM

COUJUMMB (£N3 FOCUSMO LENS

77ZZZZZZA

TOWAlV

ACCESS APERTURE 'sssssss/y/s/rsssA

Q BEAU BLOCK

TEST ARM

EXTERNAL̂ ALIGNMENT MIRROR

EXTERNAL l/j CENTERHO MIRROR

M k UK ID

FOLDING MIRROR |2

"ilUP-W ^CENTERING LENS

BEAM REIMAGMC LENS H

FUP-IH ALIGNMENT LENS

W BEAM REniACWO LENS |2

VAJOFLE ATTENUATOR

COUJMATMO LENS

rZT MIRROR

REIMAGING ARM

SEAM SHJTIEH f 2

ENTRANCE PUPfL (THEORETICAL POMT I INCH . „«. ™OU EDGE OF INTERFEROMETER) LASER SOURCE POSITION

Figure 26 LADITE Block Diagram20 o •Ct

65

REFERENCES

1. Sincerbox, G.T., "Miniature optics for optical recording", Gradient-Index Optics and Miniature Optics, Proc. SPIE 935, pg 63-76, (1988)

2. Wolniansky, P., Chase R., Rosenvold, M., Ruane, Mansuripur M., "Magneto-optical measurements of hysteresis loop and anisotropy energy constants on amorphous Tbx Fej_x alloys" J.Appl.Phys.60(l), (1986)

3. Milster, T.D., "Characteristics of Phase-compensation Techniques in Magnetooptical Read-back Systems", for publ. in Polarization Considerations for Optical Systems, SPIE, San Diego, CA (1989)

4. Finkelstein, B.I., "Laser-feedback noise in magneto-optical recording", Optical Storage Technology and Applications, Proc. SPIE 899, pg 69-76, (1988)

5. Ojima, M., Arimoto, A., "Diode laser noise at video frequencies in optical videodisc players", App. Opt. 25(9), 1404, (1986)

6. Arimoto, A., Ojima, M., "Optimum Conditions for the High Frequency Noise Reduction Method in Optical Videodisc Players", App, Opt. 25(9), 1398, (1986)

7. Keppler, and Goldman, "Analysis of Magneto-optic Media for Optical disk drives", Optical Storage technology and Applications, Proc. SPIE 899, pg 69-76 (1988)

8. Mansuripur, M., "Analysis of astigmatic focusing and push-pull tracking error signals in magnetooptical disk systems", App. Opt. 26(18), pg 3981-3986 (1987)

9. Hecht, J., The Laser Guidebook, McGraw-Hill, USA, pg 246-270 (1986)

10. Call, D.E., and Finkelstein, B.I., "Dependence of laser-feedback noise on optical-path length", Optical Data Storage Topcial Meeting SPIE Vol. 1078, pg 272-278 (1989)

11. Sharp Laser Diode User's Manual, Sharp Corporation, Osaka, Japan (1988)

12. Private communication with Blair I. Finkelstein, University of Arizona, Optical Sciences Center, (1989)

13. Private communications with Scott B. Hamilton, Apex Systems, Inc., Boulder, CO (1988)

14. Private communications with Scott DeVore, University of Arizona, Optical Sciences Center. (1989)

15. DeVore, S.L., "Simulation Methods for Optical Disk Drive Functions", Ph.D. Dissertation, University of Arizona (1988)

16. Braat, J., "Read-out of Optical Discs", in Principles of Optical Disc Systems, ed. G. Bouwhuis, Adam Hilger, Bristol (1985)

66

17. Private communication with Scott L. DeVore, Tom D. Milster, Blair I. Finkelstein, University of Arizona, Optical Sciences Center, (1989)

18. Zajaczkowski, J.S., "Proposed American National Standard, Unrecorded Optical Media Unit for Digital Information Interchange - 130mm Reversible Optical Disk Cartridge", Revision X3B1I/88-094R2, (1988)

19. Private communications with Tom Milster, University of Arizona, Optical Sciences Center. (1989)

20. Instruction Manual for LADITE Laser Diode Tester, WYKO Corporation, Tucson, AZ (1987)