1 Can we afford to build an extremely large groundbased
diffraction limited optical/IR telescope? Jim Oschmann Francois
Rigaut Mike Sheehan Larry Stepp Matt Mountain Gemini Observatory
Slide 2 2 Can we afford to build an extremely large groundbased
diffraction limited optical/IR telescope? Or can we afford ~
$1,000M Probably yes... Slide 3 3 Framework for a credible
Extremely Large/Maximum Aperture Telescope Concept Science Case An
adaptive optics solution A telescope concept A viable instrument
model Gallagher et al, Strom et al Rigaut et al Ramsay Howat et al
Mountain et al Slide 4 4 Simulated NGST K band image Blue for z = 0
- 3 Green for z = 3 - 5 Red for z = 5 - 10 = 0.1 Spectroscopic
Imaging at 10 milli-arcsecond resolution 48 arcseconds 2K x 2K IFU
0.005 pixels - using NGST as finder scope Slide 5 5 Modeled
characteristics of 20m and 50m telescope Assumed detector
characteristics m < m 5.5 m < m I d N r q e I d N r q e 0.02
e/s 4e 80% 10 e/s 30e 40% Assumed point source size (mas) 20M 1.2 m
1.6 m 2.2 m 3.8 m 4.9 m 12 m 20 m (mas) 20 20 26 41 58 142 240 50M
1.2 m 1.6 m 2.2 m 3.8 m 4.9 m 12 m 20 m (mas) 10 10 10 17 23 57 94
(Gillett & Mountain, 1998) Slide 6 6 Relative Gain of
groundbased 20m and 50m telescopes compared to NGST Groundbased
advantage NGST advantage Imaging Velocities ~30km/s Slide 7 7 An
Adaptive Optics Solution Slide 8 8 Slide 9 9 Slide 10 10 An
Adaptive Optics Solution Slide 11 11 No correction (AO off) (Rigaut
et al) New Directions for Adaptive Optics ~ arcminute corrected
FOVs possible (Rigaut et al) Numerical simulationsNumerical
simulations 5 guide stars & 5 Wavefront sensors 2 mirrors 8
turbulence layers 40 Field of view J band Fully corrected PSF
across full field of viewFully corrected PSF across full field of
view Slide 12 12 No correction (AO off) (Rigaut et al) New
Directions for Adaptive Optics ~ arcminute corrected FOVs possible
(Rigaut et al) Numerical simulationsNumerical simulations 5 guide
stars & 5 Wavefront sensors 2 mirrors 8 turbulence layers 40
Field of view J band Fully corrected PSF across full field of
viewFully corrected PSF across full field of view MCAO on Slide 13
13 No correction (AO off) (Rigaut et al) New Directions for
Adaptive Optics ~ arcminute corrected FOVs possible (Rigaut et al)
Numerical simulationsNumerical simulations 5 guide stars & 5
Wavefront sensors 2 mirrors 8 turbulence layers 40 Field of view J
band Fully corrected PSF across full field of viewFully corrected
PSF across full field of view MCAO on Optical Performance - Strehl
Ratio at 500nm across a 20 x 20 FOV (Ellerbroek,1994)
Multiconjugate Adaptive Optics On Axis Edge FOV Corner FOV 0.942
0.953 0.955 Slide 14 14 Instrumentation -- the next constraint?
(Ramsay Howatt et al) 2K x 2K IFU 0.005 pixels 10 arcsec R = 8,000
across J, H & K 4.2 x 10 9 18.5 m pixels 1.2 m Slide 15 15
Instrumentation -- the next constraint? (Ramsay Howatt et al) 2K x
2K IFU 0.005 pixels 10 arcsec R = 8,000 across J, H & K Lets
not assume diffraction limited instruments for 30m ~ 100m
telescopes will be small 6.7 X 10 7 Pixels Slide 16 16 The next
step ? 50m telescope 0 A 400 year legacy of groundbased telescopes
Slide 17 17 Technology has made telescopes far more capable, and
affordable 0 Slide 18 18 Technology has made telescopes far more
capable, and affordable Slide 19 19 Technology has made telescopes
far more capable, and affordable Slide 20 20 Optical Design
RequirementsRequirements 50m aperture Science field of view 0.5 -
1.0 arcminutes Useable field of view 1.0 - 2.0 arcminutes (for AO
tomography) Minimize number of elements (IR performance) Aim for
structural compactness KISS Slide 21 21 Optical Design 50m F/1
parabola M1, 2m diameter M2 2m diameter Slide 22 22 Optical Design
~ 3m F/20 Cassegrain focus Slide 23 23 Optical Design ~ 3m F/20
Cassegrain focus Adaptive Optics Unit Cassegrain Instrument #1
Cassegrain Instrument #2 Slide 24 24 Optical Performance 0 arcsec30
arcsec 1 arcminute FOV (Science Field) Slide 25 25 Optical
Performance 0 arcsec. 30 arcsec. 60 arcsec. Guide star FOV Slide 26
26 Optical Performance rms wavefront error 1 micron wavelength /10
0 arcsec 3060 Slide 27 27 Primary Mirror Approach Slide 28 28
Primary Mirror Approach The volume of glass in a 50-mm thick
8-meter segment is 2.5 cubic meters. This volume is equivalent to a
stack of 1.5-meter diameter boules 1.4 meters high. F/1 Segmented
Parabola Segment testing (no null lenses) 50m ~25m Slide 29 29
Primary Mirror Approach Actively controlled polishing The sag of an
8-meter segment is only 80 mmTesting Ion Figuring Final Testing
Slide 30 30 l To reduce mass, reduce mirror substrate thickness ~
50mm (1/4 of Gemini, ESO-VLT) l Individual segments still have to
be supported against self weight Primary Mirror Support Slide 31 31
Primary Mirror Support Slide 32 32 Primary Mirror Support
Gravitational print through requires between 120 - 450 support
points for a 20 cm thick meniscus Slide 33 33 Primary Mirror
Support - continued As self weight deflection D 4 /t 2, ~8m
diameter, 50mm segment will need ~ 1800 support points How many
active support points do we need to correct deformations due to
wind and thermal gradients? Slide 34 34 Primary Mirror Support -
continued As self weight deflection D 4 /t 2, ~8m diameter, 50mm
segment will need ~ 1800 support points How many active support
points do we need to correct deformations due to wind and thermal
gradients? Estimate 1 in 6, ~ 300/segment which implies > 10,000
actuators to actively support a 50m mirror Slide 35 35 Does
maintaining 10,000 actuators challenge the Quality Control
Engineers? What Mean Time Between Failures (MTBF) does this
require?What Mean Time Between Failures (MTBF) does this require?
Assume 95% up-time, over 356 x 12 hour nights Assume unacceptable
performance will occur when 5% of actuators fail Assume it takes 1
hour to replace actuator, and that we can service 8 actuators a
day, over 250 maintenance days Therefore we can replace/service
2,000 actuators/year MTBF required is 380,000 hoursMTBF required is
380,000 hours Required service life of each actuators, assuming
maintenance is 5 yearsRequired service life of each actuators,
assuming maintenance is 5 years Slide 36 36 Challenges for the
Structural Engineers... ChallengesChallenges 20mm mirror substrate
still weighs ~ 110 kg/m 2 (c.f ~ 75 kg/m 2 for Gemini/Zeiss M2)20mm
mirror substrate still weighs ~ 110 kg/m 2 (c.f ~ 75 kg/m 2 for
Gemini/Zeiss M2) Mirror segments + cells could weigh 5.5 x 45 + 200
= 450 tonnesMirror segments + cells could weigh 5.5 x 45 + 200 =
450 tonnes Wind..Wind.. 10 m/s across 50m a lot of energy at ~ 0.2
Hz10 m/s across 50m a lot of energy at ~ 0.2 Hz Telescope Optical
Structure Requirements: 50m surface must be held ~ /10 against
gravitational and wind loads 50m surface must be held ~ /10 against
gravitational and wind loads Relative pointing and tracking ~ 3
arcseconds rms Relative pointing and tracking ~ 3 arcseconds rms
Absolute pointing/tracking provided by Star-tracker Absolute
pointing/tracking provided by Star-tracker Precision
guiding/off-setting controlled by M4 and A&G/AO system
Precision guiding/off-setting controlled by M4 and A&G/AO
system Clean top-end for IR emissivity, but rigid enough to launch
5 laser beacons Clean top-end for IR emissivity, but rigid enough
to launch 5 laser beacons Slide 37 37 Challenges for the Structural
Engineers... ChallengesChallenges 20mm mirror substrate still
weighs ~ 110 kg/m 2 (c.f ~ 75 kg/m 2 for Gemini/Zeiss M2)20mm
mirror substrate still weighs ~ 110 kg/m 2 (c.f ~ 75 kg/m 2 for
Gemini/Zeiss M2) Mirror segments + cells could weigh 5.5 x 45 + 200
= 450 tonnesMirror segments + cells could weigh 5.5 x 45 + 200 =
450 tonnes Wind..Wind.. 10 m/s across 50m a lot of energy at ~ 0.2
Hz10 m/s across 50m a lot of energy at ~ 0.2 Hz Telescope Optical
Structure Requirements: 50m surface must be held ~ /10 against
gravitational and wind loads 50m surface must be held ~ /10 against
gravitational and wind loads Relative pointing and tracking ~ 3
arcseconds rms Relative pointing and tracking ~ 3 arcseconds rms
Absolute pointing/tracking provided by Star-tracker Absolute
pointing/tracking provided by Star-tracker Precision
guiding/off-setting controlled by M4 and A&G/AO system
Precision guiding/off-setting controlled by M4 and A&G/AO
system Clean top-end for IR emissivity, but rigid enough to launch
5 laser beacons Clean top-end for IR emissivity, but rigid enough
to launch 5 laser beacons Slide 38 38 Resonant Frequencies of Large
Telescopes Slide 39 39 Resonant Frequencies of Large Telescopes
Frequency (Hz) Telescope Aperture 50m 2Hz Parabolic Reflector
Antenna Systems Optics Systems (Laser/Infrared) Lowest Servo
Resonant Frequency Slide 40 40 Conceptual Design for an F/1 50m
Optical/IR Telescope Slide 41 41 Optical/Mechanical concept
Mirror-to-cell actuators Integrated mirror/cell segment Large
stroke actuators Mirror support truss with smart structure
elements/active damping as needed Three levels of figure control:
Each mirror segment Each mirror segment is controlled within an
individual cell is controlled within an individual cell Each cell
is then controlled with respect to the primary mirror support
structure Each cell is then controlled with respect to the primary
mirror support structure The support structure may have to use
smart structure technology to maintain sufficient shape and/or
damping for slewing/tracking The support structure may have to use
smart structure technology to maintain sufficient shape and/or
damping for slewing/tracking Slide 42 42 Concept Summary Optical
support structure uses at least three levels of active control
Slide 43 43 Concept Summary Adaptive Optics Unit Cassegrain
Instrument #1 Cassegrain Instrument #2 Optical support structure
uses at least three levels of active control Collimated beam allows
M3 & M4 to be tested independently and allows AO/instrument
structure to be rigidly coupled to F/20 focus - insensitive to
translation or rotation relative or rotation relative to 50m
structure to 50m structure Slide 44 44 Concept Summary Adaptive
Optics Unit Cassegrain Instrument #1 Cassegrain Instrument #2
Optical support structure uses at least three levels of active
control Collimated beam allows M3 & M4 to be tested
independently and allows AO/instrument structure to be rigidly
coupled to F/20 focus - insensitive to translation or rotation
relative or rotation relative to 50m structure to 50m structure M2
easy to make/test - may need a little more rigidity. rigidity.
Slide 45 45 An Enclosure for 50m -- how big? Restrict observing
range to airmasses < 2.0Restrict observing range to airmasses
< 2.0 30 degrees 75m Astro-dome approachAstro-dome approach 150m
75m Slide 46 46 An Enclosure for 50m -- how big? Restrict observing
range to airmasses < 2.0Restrict observing range to airmasses
< 2.0 30 degrees 75m Astro-dome approachAstro-dome approach
Heretical proposition #1 - excavateHeretical proposition #1 -
excavate significantly lowers enclosure cost further shields
telescope from wind reliant on AO to correct boundary layer 150m
75m Slide 47 47 An Enclosure for 50m -- how big? Restrict observing
range to airmasses < 2.0Restrict observing range to airmasses
< 2.0 30 degrees 75m Astro-dome approachAstro-dome approach
Heretical proposition #1 - excavateHeretical proposition #1 -
excavate significantly lowers enclosure cost further shields
telescope from wind reliant on AO to correct boundary layer 150m
75m Heretical proposition #2 - perhaps the wind characteristics of
a site are now more important than the seeing
characteristicsHeretical proposition #2 - perhaps the wind
characteristics of a site are now more important than the seeing
characteristics Slide 48 48 Framework for a credible Extremely
Large/Maximum Aperture Telescope Concept Science Case An adaptive
optics solution A telescope concept A viable instrument model Slide
49 49 Image of a 21 st Century Ground-Based Observatory -- 50m
Class Slide 50 50 Slide 51 51 How do we cost a 50m? (1999) $522
Contingency $100M Slide 52 52 How do we cost a 50m? Risk assessment
Adaptive OpticsAdaptive Optics multiple-conjugate AO needs to be
demonstrated deformable mirror technology needs to expanded for 50m
( x 10 - 20 more actuators How do we make a light-weight, 4 - 8m
aspheric segment mounted in its own active cell and can we afford
45 - 180 of them?How do we make a light-weight, 4 - 8m aspheric
segment mounted in its own active cell and can we afford 45 - 180
of them? How much dynamic range do we need to control cell- segment
to cell-segment alignment ?How much dynamic range do we need to
control cell- segment to cell-segment alignment ? Will smart,
and/or active damping systems have to be used telescope evaluate by
analysis and test. Composites or Steel? Slide 53 53 Risk assessment
- continued Telescope Structure and wind loading We need to
characterize this loading in a way that is relatively easy to use
in finite element analysis. This is easy, but mathematically
intensive. Basically for each node that gets a wind force, a full
vector of force cross spectra is generated, therefore the force
matrix is a full matrix with an order equal to the number of forces
(10s of thousands). Enclosure concept (do we need one)? What
concept can we afford both in terms of dollars/euros and
environmental impact (note Heretical Proposition #2) WE NEED A
TECHNOLOGY TEST-BED a 10m - 20m new technology telescope this is
probably to only way to establish a credible cost for a 50m - 100m
diffraction limited optical/IR groundbased telescope Slide 54 54
Supposing a tree fell down Pooh, when we were underneath it?
Supposing it didnt, said Pooh after careful thought. The House at
Pooh Corner The House at Pooh Corner Slide 55 55 Supposing we
couldnt afford a 50 or 100m Pooh, when we could have been doing
something more useful ` Supposing we could, said Pooh after careful
thought. With apologies to With apologies to The House at Pooh
Corner