Transcript Final Presentation Slides (PowerPoint Version)

Year 4 University of Birmingham Presentation
Applications of Piezo Actuators for
Space Instrument Optical Alignment
Michelle Louise Antonik
520689
Supervisor: Prof. B. Swinyard
Outline of Presentation
•
•
•
•
•
•
•
Introduction
Background
Method Developed
Results Gained
Future Possibilities
Conclusions
Questions
Introduction
• This project aims to create an alignment system
that can see movement of 1μm
Why?
• The James Webb Space Telescope will carry
several instruments that operate at cryogenic
temperatures
• RAL is tasked with observing structural changes
in one of the instruments, MIRI, at 7K
• The alignment device that is to be used currently
only has a accuracy of 6μm
Fibre Optic Sensors
• Fibre optic
sensors are to be
used to detect the
change in shape
• All electronics are kept out of the cryostat
making it a passive system
• Work by sending a light signal down the optical
fibre which is reflected off a surface back into the
fibre.
Purpose of Alignment System
• Need to increase accuracy of the fibre optic
system
• Create an alignment system that designed to
detect small movements
• Piezo actuator used to simulate the movements
of MIRI
• Piezo actuator’s movement is unknown
• Calibrate it sufficiently becomes predictable and
repeatable
Purpose of Alignment System
• Once piezo actuator is calibrated place in MIRI’s
position
• If small movements by the piezo actuator
produce repeatable changes in the fibre optic
sensor’s output then the sensors can be
calibrated to a higher accuracy
Alignment Systems
What are alignment systems?
• Alignment systems allow components to be
placed in the correct relative positions to each
other, essential for high accuracy work
• Many different types
• Interested in passive, optical systems
• These are cheaper and less noisy than other
systems
Alignment Systems
• Passive optical alignment systems are either
photogrammetry based or laser based
• Photogrammetry based systems have maximum
accuracies of 150μm
• Laser based systems using interferometry have
accuracies of less than 1nm
How the Piezo Actuator Works
What are piezo actuators?
• Is based around a piezoelectric crystal
• These crystals expand or contract when a
potential difference is placed across them
• Two types of crystal: ferroelectric and nonferroelectric
Ferroelectric Crystals
• Ferroelectric crystals have two or more stable
orientations in which the atoms can be arranged
• By applying a mechanical stress across the
crystal the atoms are forced into more compact
arrangement
• Change of ion’s position changes the polarisation
of the crystal
Courtesy of C. Kittel, Introduction to Solid State Physics, 5th ed., 1976, John Wiley & Sons Inc.
Non-Ferroelectric Crystals
• Non-ferroelectric crystals have three equal
dipole moments that have a sum at the vertex of
zero
• A mechanical stress compresses the crystal
which distorts the dipoles
• When the sum at the vertex is not zero, there is
a polarisation across the crystal
Courtesy of C. Kittel, Introduction to Solid State Physics, 5th ed., 1976, John Wiley & Sons Inc.
Piezoelectric Crystals
• Equations for crystal’s polarisation and elastic
strain both contain the stress the crystal is
under and the electric field affecting it
How the Piezo Actuator Works Cont.
a – Piezoelectric crystal, b – Sliding block, c – guiding rod,
d – fixed frame
Courtesy of attocube systems’ User Manual Inertial XYZ
Positioner ANPxyz100.
Rough Calibration of the Piezo
• Initially a rough
calibration of the
piezo was required to
understand it’s
movement
• This was done by
using a linear voltage
displacement
transducer (LVDT)
• An LVDT probe has a
central core that is
pushed into three wire
coils
Rough Calibration of the Piezo
• LVDT placed against the piezo actuator
• Piezo actuator moved outwards by 50 steps at a
time
0.030
Mean Step Size (mm)
0.025
0.020
0.015
0.010
0.005
0.000
0
2000
4000
6000
8000
Step Number
10000
12000
14000
Original Method
Fine calibration of piezo actuator
• Original idea was a basic phogrammetry
technique with simple geometry
• As the mirror moved the laser beam travelled
further
• Angles translated this to movement across the
webcam
Limitations Imposed
• Limitations on the sensitivity of the alignment
system are imposed by the equipment used
• The main limitations are
– Resolution of the webcam’s CCD
– Fish-eye lens has low sensitivity for small movements
– Angle of laser beam on the mirror
Limitations Imposed
Resolution
• The resolution is the smallest possible distance
between two points that the camera can see
• Is given by the Rayleigh Criterion:
Θ = 1.22λ/D
where λ is the wavelength of light and Θ and D are
given as below
Limitations Imposed
With the fish-eye lens
• Fish-eye lens allows large viewing area for a
small detector
– Small movements near the optical axis become hidden
• Resolution was found by moving a large light
source away from the webcam
• The height of the source was plotted against
distance from the webcam
• One standard deviation was found to be 2 pixels
• Gave resolution of 20μm
750
Height of light
source (pixels)
650
550
450
350
250
35
40
45
50
55
60
65
Distance from lens (mm)
70
75
80
85
Limitations Imposed
Without the fish-eye lens
• Resolution measured by
moving webcam
perpendicular to laser
beam
• Linear relationship
between distance moved
by webcam and laser
beam across CCD
• At regular intervals
images were taken and
brightness measured
Limitations Imposed
• One standard deviation for the points from the
line allows an accuracy of the position to be
taken to 0.5 pixels
• The resolution was 1.5μm
555
Average Centroid (pixel)
550
545
540
535
530
525
520
515
510
0.00
0.02
0.04
0.06
Position (mm)
0.08
0.10
0.12
Limitations Imposed
Angle of the laser beam
• The angle at which the laser beam hits the mirror
is the angle at which it is reflected
• A larger incident angle gives greater
magnification of the mirror’s movement
• Large angle means larger cross-section, gives
less precision
• Angles less than
20° needed
• Even at maximum
resolution is still more
than 1μm
Development of Method
• As it was not possible to use photogrammetry
techniques, interferometry techniques were tried
instead.
• A Michelson interferometer was created
• Have
accuracies of λ/2
• Any noise will
be known to come
from cryostat
rather than
alignment system
Development of Method
•
Dark rings are from destructive interference
– pd = mλ/2
• Bright rings are from constructive interference
– pd = mλ
• Movement of
piezo causes
rings to disappear
into centre
Courtesy of
www.search.com/reference/Interference
Development of Method
• Michelson set-up:
a – laser, b – beam expander, c – polariser, d – iris, e – half silvered
mirror, f – full silvered mirror, g – piezo actuator with full silvered mirror
mounted on top, h – lens, i – webcam, j – optical axis
Development of Method
• Laser light too coherent
– Small defects in the set-up obscure the results
• A less perfect light source is needed
– Gives more complicated pattern
• Use a white filtered source
Final Method
• Adapted Michelson:
b – beam expander, d – iris, e – half silvered mirror, f – full
silvered mirror, g – piezo actuator with full silvered mirror
mounted on top, h – lens, i – webcam, j – optical axis, k – white
light source with red filter, l – second iris
Results
Finding the Zero Path Difference Area
156
155
Maximum Gray Scale Value
154
153
152
151
150
149
148
147
146
0
1000
2000
3000
4000
Step Num ber
5000
6000
7000
Results
Close-up of the Zero Path Difference Area
130
Gray Scale Value
128
126
124
122
120
118
3900
4100
4300
4500
4700
4900
Step Number
5100
5300
5500
Results
• Do not see the
expected pattern
• Get zero intensity
where a peak is
expected
• Data still useable as
distance between null
points is the same as
distance between
peaks
Results
Why do you not see the expected pattern?
• Webcams are designed to view large images
• Software maybe reducing fringes as they are
small fluctuations
• The intensity at a given point due to polarisation
is
*
I  E p1 ( ) E p2 ( )
• Light maybe changing polarisation slightly during
reflection
Results
• Rough surfaces can generate a consistent phase
change
• Path difference through the half silvered mirror is
not equal
• Need to remove background from results
Results
Background Removed from Image Taken Just Outside the Zero Path
Difference Area
Gray Scale Value
10
5
0
-5
-10
-15
-20
0
50
100
150
200
Distance Along CCD
250
300
Real Fringe Pattern
• If two mirrors are not aligned exactly a fringe
pattern occurs
Courtesy of ‘Optics’ by E. Hecht
Results
Background Removed from Image Inside the Zero Path Difference
Area
60
Gray Scale Value
40
20
0
-20
-40
-60
0
50
100
150
200
Distance Along CCD
250
300
Future Work
Refine results
• Replace webcam with photo diode
• Step size of the piezo actuator
– Initial calibration found the step size to be approximately
400nm
– Wavelength of light is approx 600nm
• Can only see to 2λ/3
Future Work
• Cryogenically cool the piezo actuator for
recalibration
• Piezo steps sizes
will change as the
piezo contracts
• Replace detector
with fibre optic
sensors
Conclusion
• The Michelson interferometer gives highly
accurate results ensuring that noise detected will
come from the cryostat
• Several adaptations needed before final
calibration
– Replacing the webcam
– Adding in a compensator plate for the optical path
difference
– Reducing piezo actuator’s step size
Questions