Transcript Slides - Caltech Optical Observatories

A Short Introduction
to Adaptive Optics
Presentation for NGAO Controls Team
Erik Johansson
August 28, 2008
Overview
•
•
•
•
•
•
•
•
•
Why we need AO
The basics of AO
Intro to wavefront sensing
Intro to tip-tilt correction
Intro to higher-order wavefront correction
LGS vs NGS AO
Limitations of AO
How NGAO will differ form our current AO system
Q&A
2
Why do we need AO?
Short exposure images of the stars Gamma Perseus
and Alpha Orionis (Betelgeuse) demonstrate the
effects of atmospheric turbulence
3
Without atmosphere, the telescope forms a
perfect “diffraction-limited” spot in the focal
plane
Light from distant star
Telescope aperture
Focal Plane
Image
Spot size = 2.44 l/D
4
The atmosphere acts like many lenses of size r0
to create moving “speckles” in the image
Light from distant star
Atmosphere (lens size = r0)
Telescope aperture
Focal Plane
Freeze the speckles by using short
exposures < ~0.1 sec
Image
r0 is characteristic size of the
atmosphere
Number of speckles ~ (D/r0)2
Spot size = 2.44 l/r0
First characterized by Fried in 1966
What is D/r0 for Keck?
5
A broad optical bandwidth smears the
speckles out in a radial fashion
Narrowband
Broadband
(Credit C. Neyman, AMOS)
6
Details of diffraction from
circular aperture
1) Amplitude
First zero at
r = 1.22 l / D
2) Intensity
FWHM
l/D
7
8
Diffraction pattern from
hexagonal Keck telescope
Stars at Galactic Center
Ghez: Keck laser guide star AO
9
What is a spatial frequency?
A sheet with a sinusoidal “wave” which can vary in
frequency (wavelength) and orientation (direction)
A spatial frequency also has phase: its peaks
and valleys have some kind of reference to a
known point in the image
10
How does the atmosphere affect system
performance?
Telescope OTF
Seeing Limited TF
Tip-Tilt Compensated TF
For D/r0 = 15
Normalized Spatial Frequency
11
The Basics of AO
How does AO work?
AO corrects distorted wavefronts in real time to
compensate for blurring effects of the atmosphere
13
What do AO and flying potato
chips have in common?
14
Intro to Wavefront Sensing
How do we measure wavefronts?
Detectors cannot measure the phase of the
light, only the intensity.
16
17
Intro to Tip-Tilt Correction
Tip-tilt correction
Mirror
Disturbance
Vector
Tip-Tilt Mirror Controller
Tip-Tilt
Sensor
Residual
Tip-Tilt
Error
(arc-sec)
S
Control law
Servo
S
Rotation
(UTT only)
Mirror
Position
Commands
(arc-sec)
Closed-loop
Mirror
Positioning
Controller
(CLMP)
Telemetry
Recorder
(TRS)
Angular Offset
(DT Ctrl Offset)
Control law
Parameters
Loop cmd
Mirror
Offset
Variable Rotation
Angle
19
Closed Loop Mirror Positioning
Closed-Loop Mirror Positioning Controller
Atmospheric
Tip-Tilt
Controller
Mirror
Position
Commands
(arc-sec)
Arc-sec to
actuator space
conversion
Digital to
Analog
Converter
PID
Servo
Current
Mirror
Position
Conversion
Matrix
Bridge
Sensors
High voltage
Amplifier
High voltage
Actuator
Signals
Mirror
Actuators
Strain
Gauge
Outputs
Servo
Parameters
20
Intro to Wavefront
Reconstruction and Correction
22
23
WFC main data flow
Raw frames
Centroids
Subap intensity
Wave
Wave Front
Sensor
(WFS)
Camera
Front
Residual WF error
RSS Residual WF Error
Tip-tilt error
WFS focus error
Processor
Actuator vector
DM focus
(WFP)
Background
Compensation
WFS
HW
IF
Flat Field
Compensation
Centroid
Computation
Telemetry
Recorder
(TRS)
MatrixVector
Multiply
DM
HW
IF
Deformable
Mirror
(DM)
Tip-tilt
Tip-tilt
Controllers
(DTT/UTT)
Control law
Servo
Pixel threshold
error
WFS
parameters
Background image
Flat field
Intensity threshold
Centroid gain
Centroid origin
Reconstruction
Matrix
Control law
Parameters
Loop command
Actuator map
DM origin
24
How a deformable mirror works
(idealization)
BEFORE
Incoming
Wave with
Aberration
Deformable
Mirror
AFTER
Corrected
Wavefront
25
Deformable Mirror for real
wavefronts
26
Most deformable mirrors today have thin
glass face-sheets
Glass face-sheet
Light
Cables leading to
mirror’s power
supply (where
voltage is applied)
PZT or PMN actuators:
get longer and shorter
as voltage is changed
Anti-reflection coating
27
Front view of Xinetics DM (Keck)
349 degrees
of freedom;
~250 in use
at any one
time
(paper
coasters)
28
What are MEMs deformable mirrors?
A promising new class of
deformable mirrors, called
MEMs DMs, has emerged in
the past few years.
Devices fabricated using
semiconductor batch
processing technology and
low power electrostatic
actuation.
Potential to be very
inexpensive ($10/actuator
instead of $1000/actuator)
MEMS:
micro
micro-electro
electro-mechanical
systems
Boston University MEMS Concept
Electrostatically
actuated
diaphragm
Attachment
post
Membrane
mirror
Continuous mirror
Boston University
Boston MicroMachines
• Fabrication: Silicon
micromachining
(structural silicon and
sacrificial oxide)
• Actuation: Electrostatic
parallel plates
NGS vs LGS AO
NGS AO Control
Light from
Telescope
Telescope pointing offload
Tip/tilt
NGS
IR transmissive
dichroic
Science
Camera
Offload
focus to
telescope
Wavefront
Controller
beamSplitter
WFS
NGS Reconstructor
Centroid Origins
Flux
Rot & pupil angle
When TT closed
LGS AO Control
Light from
Telescope
Telescope pointing offload
Tip/tilt
LGS
NGS
IR transmissive
dichroic
Science
Camera
Offload
focus to
telescope
Sodium
transmissive
dichroic
Wavefront
Controller
STRAP
Lenslets
WFS
LBWFS
Focus
Optimized centroids offsets
Tip/tilt
to Laser
LGS Reconstructor
Laser
TT mirror
TSS x,y,z stage
Laser Orientation
Spot size & flux
Rot & pupil angle
When DM closed
Laser pointing offload
Limitations of AO
• Isoplanatism
– Tip-Tilt Isoplanatism
– Focus isoplanatism
• Sky coverage
– WFS sensitivity
– TT sensor sensitivity
• Imaging wavelength
• Controller bandwidth
• Error budgets, and more…
34
What determines how close the
reference star has to be?
Reference Star
Turbulence has to be
Science
Object
similar on path to
reference star and to
science object
Common path has to be
large
Turbulence
z
Anisoplanatism sets a
limit to distance of
reference star from the
science object
Common
Atmospheric
Path
Telescope
Expression for isoplanatic angle 0
• Strehl = 0.38 at  = 0
0 is isoplanatic angle
3 / 5



2
8/3
2
5/3
 0  2.914 k (sec  )  dz CN (z) z 


0
0 is weighted by high-altitude
turbulence (z5/3)
• If turbulence is only at low altitude,
overlap is very high.
Common
Path
• If there is strong turbulence at high
altitude, not much is in common path
Telescope
Anisoplanatism: how does AO image degrade as
you move farther from guide star?
credit: R. Dekany, Caltech
• Composite J, H, K band image, 30 second exposure in each
band
• Field of view is 40”x40” (at 0.04 arc sec/pixel)
• On-axis K-band Strehl ~ 40%, falling to 25% at field corner
37
AO image of sun
in visible light:
11 second
exposure
Fair Seeing
Poor high altitude
conditions
From T. Rimmele
38
AO image of sun in
visible light:
11 second
exposure
Good seeing
Good high altitude
conditions
From T. Rimmele
39
Focus Anisoplanatism:The laser doesn’t
sample all the turbulence
40
Additional slides from Claire
Max’s UCSC Class
NGWFC Results
Successes: Old vs. new
Some of the best images of a 7th magnitude star taken with
the old WFC (left) and the NGWFC (right). The images
have K-band Strehls of 58% and 66% respectively.
Strehl record: 71% at K-band
Limiting magnitude: R=16
43
NGS performance exceeds expectations
60+% Strehl for R=14 guide star
Requirement was to meet or exceed 30% Strehl for 14th
magnitude guide star in good seeing (r0 ≥ 20 cm).
Strehl record: 71% at K-band
Limiting magnitude: R=16
44
LGS performance has improved as well
• LGS AO results during especially good seeing.
• Best performance increased from 44% to 51% Strehl in K.
• Limiting magnitude R=19
45
Improved performance on Brown Dwarfs
J-band image of a brown dwarf binary pair with separation of
80 mas (Michael Liu, 26 March 2007).
46
Best LGS AO images of the galactic
center
K-band image of the Galactic Center in LGS AO (left) and NGS AO (right).
Credit: Andrea Ghez, Jessica Lu.
47
Extended Objects
• J, H and K’ color composite o Uranus (left). The inset on the
top left is an enlarged image of Miranda at K’.
• H and K’ color composite of Neptune (middle)
• K’ image of Titan (right).
48
Uranus ring crossing
The rings of Uranus as observed with the Keck AO system since 2004.
Optically-thick rings like  disappear due to inter-particle shadowing;
optically-thin rings like  brighten. Credit: Imke de Pater.
49
Improved NGS/LGS crossover point
LGS perf
NGS perf
• We are now able to use NGS in observing scenarios where
we used LGS before and get better performance
50