GSMT AO
Simulations
François Rigaut, Gemini Observatory
GSMT SWG Meeting, LAX, 2003/03/06
Presentation map
Ground Conjugated AO:
GCAO/AO/MCAO relative merits
Basic limitations and principles
Performance (Monte-Carlo & Analytical)
Astrometry with MCAO
2
GSMT SWG Meeting, LAX, 2003/03/16
Ground Conjugated
Adaptive Optics for
the GSMT
3
GSMT SWG Meeting, LAX, 2003/03/16
Ground Conjugated AO
Idea very similar to MCAO, but using a
single ground conjugated deformable
mirror instead.
The ground layer is the strongest.
The FWHM gains are modest, but the
field of view much larger than AO/MCAO
➜ tailored to different applications.
Performance independent of telescope
D
Borderline to use NGS, but LGS
available
4
GSMT SWG Meeting, LAX, 2003/03/16
relative merit
GCAO/AO/MCAO
#photons obj comp.
×
SNR gain ≈#photons
obj uncomp.
FWHM obj uncomp.
FWHM obj comp.
≈ Strehlcomp x 0.7 x D/r0
Metrics (arbitrary?):
1→ SNR gain x FoV/FoV(AO)
2 → SNR gain x [FoV/FoV(AO)] / cost
5
GSMT SWG Meeting, LAX, 2003/03/16
relative merit
GCAO/AO/MCAO
8m,1μm
AO
MCAO
GCAO
Seeing
SNR gain
5
4
2
1
FoV
20’’
80’’
10’
∞
Cost (AU)
1
3
2
-
Metric1
5
56
1800
∞
Metric2
5
19
900
∞
6
GSMT SWG Meeting, LAX, 2003/03/16
relative merit
GCAO/AO/MCAO
30m,1μm
AO
MCAO
GCAO
Seeing
SNR gain
5
4
2
1
FoV
20’’
80’’
10’
∞
Cost (AU)
1
3
2
-
Metric1
19
210
1800
∞
Metric2
5
19
900
∞
7
GSMT SWG Meeting, LAX, 2003/03/16
Geometry of the Problem
8
GSMT SWG Meeting, LAX, 2003/03/16
Geometry of the Problem
Field of View
•
The effective
thickness of the
compensated layer
depends on:
seeing
wavelength (via
d)
Field of view
→ heff ≃ 2 r
0
/θ
for good seeing:
h
heff(K) =
d
1600m
heff(V) =
Actuator grid
275m
9
GSMT SWG Meeting, LAX, 2003/03/16
Actuator Density
(Analytical approach)
Average FWHM at V
band versus the
diameter of the field of
view for actuator
densities of
Natural V band seeing
25cm (lower solid)
50cm (dot)
100cm (dash)
Pachon Cn2 profile & r0
Pachon Cn2 Profile
10
GSMT SWG Meeting, LAX, 2003/03/16
Wavelength dependency
(Analytical approach)
Natural (straight upper
lines) and
Compensated (lower
curves) FWHM
averaged across field
of view:
Bands:
V (solid),
J (dot)
K (dash)
Pachon Cn2 Profile
11
Cerro Pachon Cn2
profile
Actuator density = 1m.
GSMT SWG Meeting, LAX, 2003/03/16
Monte-Carlo Simulations
Developed a new AO simulation code (Yorick based)
Handles any number of DMs (PZT/bimorph) at any
altitude, any number of WFSs (SH/Curvature) at
arbitrary location in the field. Cone effect included.
Only simple least square implemented (upgrade).
Include most of AO effects (fitting, aliasing, servo-lag,
anisoplanatism, noise)
Goal here is to validate the analytical results and
investigate:
dependence upon telescope diameter
dependence upon number of guide stars
dependence upon compensated field of view
dependence upon the wavelength, Cn2 profile, etc...
Uniformity and shape of the PSF across the field
12
GSMT SWG Meeting, LAX, 2003/03/16
Performance vs Field of
View
(Monte-Carlo Simulations)
•
Conditions:
λ = 1.25 μm
Seeing(500nm)=0.5”
Pachon profile (64%
@ ground level)
8-m telescope
(⇔30m)
50cm pitch
10 WFSs
GS placed on a circle
No AO results:
FWHM = (400± 14)
mas
Field of View Diameter [arcmin]
13
Checked telescope
diameter has little
GSMTimpact
SWG Meeting, LAX, 2003/03/16
Performance vs Field of
View
(Monte-Carlo Simulations)
•
Conditions:
λ = 1.25 μm
Seeing(500nm)=0.5”
Pachon profile (64%
@ ground level)
8-m telescope
(⇔30m)
50cm pitch
GS placed on a circle
No AO results:
FWHM = (400± 14)
mas
Number of Guide Stars
14
GSMT SWG Meeting, LAX, 2003/03/16
Astrometry with
MCAO on GSMT
15
GSMT SWG Meeting, LAX, 2003/03/16
Astrometry with MCAO
Bears the promise of 0.1mas astrometric
accuracy (FWHM/30). But...
Difficulty intimately linked to the ability of
compensating high altitude turbulence
Not a problem with GCAO or AO
Difficult to analyze. Require a careful
consideration of biases
Established a preliminary error budget
Key: design a system as little bias-prone
as possible
16
GSMT SWG Meeting, LAX, 2003/03/16
Error budget: Assumptions
relative astrometry
3+ tip-tilt GSs adequately located in FoV
Astrometric references adequately
located in the field of view (QSOs,
background stars) used to establish
astrometric referential from observation
to observation
TT/plate scale mode sensing done at the
science detector
Errors given at the Tip-Tilt GSs
17
GSMT SWG Meeting, LAX, 2003/03/16
Astrometry with MCAO
Source
Error (mas)
Distortion calibration (high order NCPA)
0.2
Pixel QE inhomogeneities
Residual local image motion
(convergence)
WFSs calibration/Zero point
0.08
Spatial Aliasing on TipTilt GS WFSs
0.1
reduce by dithering
reduce by dithering (1.5
mas/pix)
long total exptime ≫ 1mn
0.5?
T.B. detailled
-
Beam position error of LGSs
Comment
< 0.1
astrometric reference in FoV
careful optical design
Non-linear effects in WFS/DMs
0.6
place holder
Total
0.82
needs more work
Some of these items can be checked with current systems. I suggest a plan be
prepared to get observing time on Altair or other systems.
18
GSMT SWG Meeting, LAX, 2003/03/16
Conclusions
Ground-Conjugated:
Provides factors of 2 improvement on 520’ field of view
Good PSF shape and homogeneity
Monte-Carlo simulations confirm
analytical derivations
Astrometry: A preliminary error budget
was developed. Need input from
community (astrometry and AO) to
investigate further. Possible tests on
existing systems
19
GSMT SWG Meeting, LAX, 2003/03/16
The End
...before you ask the question: This presentation was
prepared and played using Keynote, the new
presentation software from Apple.