Keck Precision AO (KPAO)

1
Keck Precision AO (KPAO)

• Notes for AOWG telecom
• July 22, 2005
• Ralf Flicker, Olivier Lai Christopher Neyman
• for the KPAO team

2
Presentation Outline

• KPAO animation
• AOWG history
• Science Case
• KPAO simulation status
• Review of current AO architectures

3
KPAO Keck Precision AO

• KPAO was identified by the AOWG as their highest
  long-term priority in support of the
  Observatorys Strategic Plan
• A very high Strehl system with stable image
  quality good sky coverage (LGS AO)
• Current effort Produce a conceptual design by
  early 2006
• Performance simulations, detailed science case
  development draft designs in process
• Will then seek external funding
• The following galactic center images (from the
  current Keck AO system at 3.5 microns wavelength)
  illustrate KPAOs intended capability at shorter
  wavelengths. In reality the resolution would
  also improve (linearly with wavelength).

4
AOWG Strategic Planning

• Proposed timeline defined at 11/02 Strategic
  Planning (SP) meeting.
• K1 laser added in early 03
• Priorities confirmed at 9/04 SP meeting.
• Mike Brown AOWG vision is that high Strehl,
  single-object, AO will be the most important
  competitive point for Keck AO in the next
  decade.
• Based on science cases presented/discussed by
  AOWG members
• Top-level requirements defined by a subgroup
  approved by AOWG

NSF-funded laser K1
Proposed Time line 11/02
14 more years to go to reach all AO all the time
5
KPAO Technical Requirements (KAON 237)

• High IR Strehls -gt evaluate 120 180 nm rms
• On-axis LGS NGS, median seeing, El gt 45?, NGS
  mv lt 17
• High Strehl stability -gt 15 nm
• Moderate field of view -gt 30 radius
• Near complete sky coverage
• Good knowledge of the delivered PSF
• Wavelength coverage -gt 0.45 to 14 ?m
• Facility-class system
• Sensitivity angular resolution not yet addressed

6
Science case

• Science case draft has been completed by the
  Olivier Lai with input from the AOWG (See KAON
  331).
• Olivier presented science case to Keck SSC
  6/29/05
• Work needs to continue in the area of flowing
  down these scientific goals into system and
  subsystem requirements.
• This work will go forward with the input of the
  new AOWG and KPAO science team.

7
KPAO performance predictions

• Flicker and Neyman has been working on analysis
  and computer simulations that can predict the
  performance of the KPAO system
• Evaluated wavefront error budgets that total 120
  nm and 180 nm respectively
• The 120 nm is a combination of analytical and
  Monte Carlo models.
• 180 nm is from Monte Carlo simulation only
• Basic AO system parameters for a narrow field AO
  (NFAO) version of KPAO have been determined

8
KPAO Monte Carlo simulation

• The next slide is a sample screen shot of the
  KPAO Monte Carlo simulation
• Based on F. Rigauts AO simulation (YAO),
  Modified by Flicker for KPAO
• Basic KPAO simulation includes
• atmospheric phase screens
• geometric optics model for wavefront propagation
• Physical optics propagation model for Shack
  Hartmann wavefront sensor
• Closed loop operation, photon and read noise.
• LGS spot elongation, Rayleigh backscatter, LGS
  centroid gain optimization
• Evaluate the PSF at user selectable locations
  over field of view

9

• WFS display (5 LGS)

• Instantaneous PSF

• Long exposure PSF

• DM shape

• Strehl History

• Residual Wavefront

10
Keck specific simulation inputs

• Simulation infrastructure exists to handle
  telescope optical aberration and vibration
  effects
• Input as temporal and spatially varying phase
  screen for now. Integrated Modeling in the future
  (??)
• Need to verify
• the inputs to the simulation
• Some test inputs to simulation shown below

Segment phase errors
Vibration spectrum
11
Large part of KPAO cost will be lasers

• Need to understand Watts/ tradeoffs
• Obtained several year data set of Na Column
  densities for Maui (http//eoslserver.csl.uiuc.edu
  /Research/Maui/NaLidar/index.html)
• Model for photon return from Na laser guide
  stars, included laser saturation effects (Both
  Pulsed and CW formats)
• Models compared to Keck/Gemini simultaneous LGS
  propagation on May 28 2005
• Comparison between models and theory good to
  about 0.4 magnitudes
• actual Na density is large uncertainty in
  verifying models

12
Sky Coverage

• Important to finalize requirement as it drives AO
  architecture (NFAO vs. MOAO/MCAO)
• Large sky coverage will require
• IR tracking detectors
• Some sharpening of NGS
• Need to measure 3-4 NGS (tip/tilt only) or 1 NGS
  (tip/tilt focus/astigmatism)
• Using tools developed by Richard Clare for TMT to
  evaluate KPAO sky coverage

13
Explanation of KPAO error budget

• Atmospheric parameters from KAON 303
• rescaled to r0 of 16 cm at 0.5 mm.
• Error allocation for Telescope, Instrument and
  Tracking using best estimates.
• 180 nm 5 LGS simple SVD reconstructor with close
  loop constraints
• 120 nm estimate the focus anisoplanatism and
  wavefront reconstruction error are estimated from
  an analytical AO model
• Working on implementing better tomography
  algorithms

14
Sample KPAO Wavefront Error Budget
Error Type Simulation parameters Simulation parameters
Fitting Error 65 54 700 act (32x32) 1300 act (40x40)
Servo Error 45 35 1000 Hz 1500 Hz
Measurement Error (noise) 60 40 15 W (CW) laser, 6e- CCD 20 W (CW), 1e- CCD
Focus Anisoplanatism 58 45 5 LGS (corners center) 7 LGS (hex. center)
Alias, reconstruction 48 35 SVD Estimate from Linear AO model
Total for Higher Order AO 125 95

Telescope 70 50 Allocation Allocation
Instrument 50 35 Allocation Allocation
Tracking (noise, servo iso.) 97 40 Allocation Allocation

Total 180 120
15
Summary

• Important to develop science requirements as part
  of the conceptual design process.
• Therefore desire more community involvement.
• Performance analysis tools in place to do
  science/technical trade studies.
• Emphasis soon to switch to system design issues.

16
Thoughts on various AO architecturesConventional
LGS/NGS AO

• What we have today
• Current NGS systems have 200 nm rms wavefront
  error, but only around bright guide stars (V13)
• Current LGS system have 300 nm rms wavefront
  error
• At a good site the cone effect error is 150 nm
• This imposes a fundamental lower limit to single
  LGS AO correction between 150-250 nm
• 250 (2002 1502)1/2

17
Thoughts on various AO architecturesExtreme AO

• One solution it to decide that NGS are the only
  viable reference source
• A high Strehl system is still possible (i.e.
  XAOPI)
• But now the number of targets is reduced to the
  brightest stars (V9 or brighter)
• This is perfectly acceptable for certain types of
  very compelling observations
• Extra-solar planets

18
Thoughts on various AO architecturesTomography

• Using multiple LGS to reduce cone effect was
  proposed in the original LGS paper (1985)
• Several variations have been proposed in recent
  years
• Multi conjugate AO (MCAO)
• Multi object AO (MOAO)
• Narrow Field AO (NFAO)
• Ground Layer AO (GLAO)
• These concepts all rely on measuring the volume
  turbulence above the telescope with several LGS
• The multi LGS measurement process has been named
  tomography in analogy with medical imaging

19
Thoughts on various AO architecturesGLAO

• Measure volume turbulence
• select out only the turbulence at the ground and
  boundary layer ( first km)
• Apply this correction with a single corrector at
  the telescope pupil
• Field of view is large because turbulence
  correction at pupil is approximately the same for
  all field angles
• wavefront error 500 nm rms

20
Thoughts on various AO architecturesMCAO

• Measure volume turbulence with several LGS
• Approximate turbulence as occurring at several
  discrete altitudes
• Apply this correction with correcting elements
  that are conjugate to each altitude (from step
  above)
• Field of view is large because turbulence
  correction is applied at altitude
• Cone effect reduced as well
• wavefront error 100-200 nm rms (field average)
• Fundamental limits set by tomography, number of
  correctors and need for NGS tip/tilt stars

21
Thoughts on various AO architecturesMOAO

• Measure volume turbulence with several LGS
• Select preferred directions for correction, i.e.
  science objects and possibly tip/tilt NGS
• Apply each direction optimized correction with a
  optic that is unique to that direction (field
  separated)
• Correction is only applied at interesting
  locations in a large field of view dont correct
  blank sky!
• As currently conceived LGS AO light is open loop
• This requires linear AO correctors and wavefront
  sensing over a large dynamic range
• Required technology has yet to be
  demonstrated/tested
• UCSC is doing conceptual design of MOAO for TMT
• Fundamental limits set by tomography, non
  linearity of AO system elements and need for NGS
  tip/tilt stars

22
Thoughts on various AO architecturesNFAO

• Measure volume turbulence with several LGS
• Select correction needed for on axis objects
• Corrected field of view small
• System can run closed loop around LGS with proper
  reconstruction matrix
• System upgrade to full MCAO straightforward
• Wavefront error 50-120 nm rms
• Fundamental limits set by tomography and need for
  NGS tip/tilt stars