This Ground Layer Adaptive Optics (AO) aims at concentrating the energy
of a Point Spread Function over a large FOV (1’) for a visible-light integral
field spectrograph: Multi Unit Spectrographic Explorer (MUSE), a second generation
instrument for the Very Large Telescope (VLT).
The prime goal of MUSE is to provide a deep 3D survey capability in
the optical range at the VLT, especially for the detection of very distant
(up to z~7) Lyman α emitters. Emphasis is put on getting a relatively wide
field with a good sky sampling. Adaptive Optics system is to get significantly
better energy concentration.
The baseline of the MUSE Ground Layer Adaptive Optics (GLAO) is to provide
two modes of correction:
• Wide field of view mode (WFM):
seeing reducer over a 1' FOV at 750nm with a factor 2 improvement in 0.2”
ensquared energywith high sensitivity and high sky coverage
• Narrow Field of view mode (NFM):
5-10” FOV with 10% Strehl ratio at 650nm.
Two AO concepts are investigated:
• GLAO#1 based on the availability/feasibility
of an adaptive secondary at VLT
• GLAO#2 based on a more classical
AO approach using relay optics and ~1k actuators deformable mirror.
Both concepts make use of 4 Laser Guide Star wavefront tomography and
one deformable mirror conjugated to the ground. Commands are optimized to
provide the best possible correction within the 1’ FOV in a so-called ground
layer correction approach. Tip-tilt correction is performed either with
an off-axis NGS in the visible (WFM) or an on-axis NGS in the NIR (NFM).
GLAO#1 and 2 have a number of design commonalities:
- GLAO correction for MUSE will be provided at the Nasmyth focus
of one VLT
- Four reconfigurable Sodium Laser Guide Stars arranged in a square
configuration and emitted from four 50 cm laser projectors located on the
VLT centrepiece ring. The four lasers are emitted toward the centre of
the scientific FOV such that to avoid Rayleigh scattered light crossing
the visible scientific FOV
- Four rotating WFSs with automatic refocusing (80-180 km)
- For the Wide Field Mode the LGS are located at 60” off-axis. For
the Narrow Field Mode the LGSs are located at 15” off-axis.
- Faint visible (mv~19) Tip-Tilt Natural Guide star sensor patrolling
a 3' FOV will provide tip-tilt correction for the Wide Field Mode.
- Faint NIR on-axis tip-tilt Natural Guide Star sensor will provide
tip-tilt correction for the Narrow Field Mode.
In GLAO#1, the Adaptive Secondary provides a corrected F/15 beam,
the LGS focusing of the WFSs is performed by 2 plates:
- Plate 1 or collecting plate:
This plate is a 16.4 m focal length lens with 4 small flat mirrors at 90°.
The collecting plate is rotating like the telescope pupil. The small mirrors
are in the telescope central obstructions of the laser beams. They collect
the focused images of the lasers and send them to the wavefront sensors.
- Plate 2 or focusing plate:
This plate (in the Wide Field Mode) is a Sodium Notch mirror transmitting
the scientific and the technical fields and reflecting the laser light towards
the collecting plate. A translation of the plate adjusts the laser focus
on the flat mirrors of the collecting plate. In the Narrow Field Mode the
focusing plate consists of four small mirror periscopes moving the 4 laser
spots from 15” to 60” off-axis such that the collecting plate remains unchanged.
These periscopes rotate like the telescope pupil. The laser beams are then
sent to 4 classical Shack-Hartman wavefront sensors.
In the GLAO#2 design, the AO relay optics is compatible with the implementation
of two deformable mirrors (conjugated at 0 & 8 km) and one separate
Tip-Tilt mirror. The ground layer deformable mirror diameter (pupil) is 180
mm. After the adaptor flange a
total
reflection prisms, with a pupil transfer lens glued on its entrance face,
directs the light down through a hole on the optical bench. An
AO optical relay is inserted between
the adaptor flange and the Nasmyth focus. The collecting plate and the wavefront
sensors have been moved outside the telescope beam and are now parallel
to the optical bench.
The
folding mirror (up) and
the 8.5 km layer conjugate mirror define the vertical axis on the platform.
After the
optical relay system,
operating at magnification one through a 3 lens system, the light reaches
the
new collecting plate which is
again a lens with a focal length of 2.8m. A flat mirror located on the back
of the prism is used to locate the focus given by the relay optics at the
nominal Nasmyth focus. The focusing plate is identical to GLAO#1. The 4 wavefront
sensors and the collecting plate have to rotate around a vertical axis to
follow the telescope pupil rotation. The focusing plate is not rotating in
the Wide Field Mode and is rotating in the case of the Narrow Field Mode.
The wavefront sensors (for both GLAO#1 and #2) are Shack Hartmann with 32x32
subapertures.
The systems consists of two parts called instrument structure
1 and 2. The AO instrument structure 1 is attached to the VLT Adapter rotator.
In
GLAO#1, the four WFSs are attached
to a mechanical structure rotating with the adapter and the WFS detectors
are mounted via an intermediate housing to the
AO structure 1.
In
GLAO#2,
Structure 1 is not rotating and
contains the relay optics, the deformable mirror conjugated at the ground,
the tip-tilt mirror the AO focal plane calibration unit and the provision
for a second deformable mirror. The tracking for the WFS Detectors is realized
internally by a rotation stage.
The
AO Structure 2 and
the implemented subassemblies are similar for both concepts. The Focusing
Plates, Calibration Plate and the MUSE Calibration Folding Mirror are located
close to the Nasmyth focus. The acquisition camera and the IR tip-tilt sensor,
including the pick up mirror unit are located after the Nasmyth focal plane.
On axis a folding mirror, with a clear aperture for the scientific path redirects
the light towards the NGS Guide Camera.
Description
|
Requirements
|
Remarks
|
Number of LGSs
|
4 CW Sodium LGSs (589 nm)
|
|
Location of emitters
|
Telescope centre piece; converging toward FOV
|
Rayleigh scattering in vis. science path
|
Return flux @ Nasmyth
|
2.5 106 Na photons/m2/s
(goal 5 106)
|
Density of Na 2 1013m2
|
Maximum spot size
|
1.25” FWHM at 45 zenith angle
Seeing 0.8”
|
|
LGS pointing precision
|
1” goal 0.5”
|
|
Jitter control
|
Refresh rate 1kHz, residual <50mas
|
|
The
Real-Time computer will
receive a dataflow of four times 32x32 sub-apertures with 6x6 pixels (192x192
pixels). The Real-Time Computer of the MUSE-AO system is based on the ESO
RTC platform concept.
In the WFOV mode, we take advantage of the reduced requirements for
the final Strehl ratio and the fact that correcting for a large field-of-view
implies a significant average of the measured slopes. Each WFS detector
will be managed by a different controller and the output of the detector
routed to a different computational board. Each of these boards will perform
the classical operations of:
- data descrambling and conversion from integer to floating point
numbers
- flat fielding and background subtraction
- slope computation and reference subtraction
This parallel process completes in about 100µs. Data are then
sent to the reconstruction module. The main design decision takes place
at this point where the slopes computed independently for each sensor are
then sent to the same back-end computational module where they are averaged
and applied to a single small control matrix to project them into the mirror
voltage space, thus saving a considerable amount of computing power.
As back-end the system features another board to collect all the data
plus the information from the tip/tilt sensor to implement the controller
and to control the deformable mirror. Another board receives all the data
from the reconstructor to perform statistics. Optionally, it can also receive
data from the control module if a PMC extender is used to add another link
card. Without considering pipelining and considering a communication time
of 100µs for each hop, the total computational time will be about 500µs.
So a frame rate of 700Hz can be achieved with a load of 35% and 1.0 KHz
can be achieved with a load of 50%.
In the NFOV mode we cannot average the slopes directly after their computation
and we need to go through the complete reconstruction of the wavefront in
each of the four sensors. We can do it in a modular way designing the system
as the pipeline of the same acquisition module as proposed before plus
one reconstructor dedicated to each sensor. A back-end stage collects the
data from the 4 reconstructors to implement the controller. A statistical
board is also present to compute statistics on the real-time data. To minimize
the connections between the various boards a high degree of pipelining must
be used.
MUSE and MUSE-AO
Software design
is an ESO standard architecture.
Due to the rayleigh scattered light constrains, the LGSs are positioned
60’’ off-axis for the Wide FOV mode. For the Narrow FOV mode the 4 LGSs are
located at 15" off-axis. Simulation parameters for both the wide and Narrow
FOV are reported in the table below.
Parameters
|
Values WFOV mode
|
Values NFOV mode
|
Number of sub-apertures / LGSs
|
32x32 / 4
|
32x32 /4
|
LGS position & flux
|
60" off-axis / 150 ph/ sub-aperture / s
|
15" /350 ph/ sub-aperture / s
|
Frame rate & temporal delay
|
700 Hz/ 2 frames
|
700 Hz/ 2 frames
|
Loop gain
|
0.4
|
0.5
|
Tip-tilt NGS flux & Position
|
150 ph / sub-aperture / s / 75’’ off-axis
|
1000 ph / sub-aperture / s / on-axis
|
Wavelength
|
0.75 µm
|
0.75 µm
|
Seeing/ Tau0
|
1.1’’/ 3ms
|
0.65" / 3ms
|
Number of active actuators/ subapertures
|
881 /768
|
881/768
|
The
performance of the Wide Field
Mode is given as a function of pixel size on-axis and at 30“off-axis.
The bottom curve provides the encircled energy of an uncorrected PSF. The
WFOV AO system needs a suitable NGS for tip-tilt correction.With a Tip-tilt
NGS limiting magnitude of 17.5 and a search FOV of 3’, more than
70% of selected deep fields can be observed
The
Strehl ratio and
FWHM performances of the Narrow
Field Mode FWHM are shown as a function of wavelength. The instrumental
error budget as well as LGS spot elongation and Na height variations have
been neglected here (preliminary estimates are 80 nm rms equivalent to Sr(650)
loss of 55 %). Tip-tilt measurement is done on-axis on the NIR counter part
of the observed object.
- F. Henault, R. Bacon et al. MUSE, “a second generation integral field
spectrograph for the VLT”, SPIE 5492, 2004
- M. Le Louarn, N. Hubin, “Wide-field adaptive optics for deep-field
spectroscopy in the visible”, Monthly Notices of the Royal Astronomical
Society, Vol. 349, Issue 3, 2004.
- Le Louarn, N. Hubin, R. Arsenault, “Adaptive Optics for second generation
VLT instruments” SPIE 5490, 2004
- R. Stuik et al. , “generalised sky coverage for Adaptive Optics
and interferometry”, SPIE 5490, 2004
![[ More info about ESO ]](eso-logo_002.gif){kind=logo}
