The onsky NGSLGS MOAO demonstrator for EAGLE

1
The on-sky NGS/LGS MOAO demonstrator for EAGLE

• Tim Morris
• Durham University

2
Talk overview

• MOAO with EAGLE
• CANARY concept
• Optomechanical design
• Subsystem performance
• System performance
• System calibration tasks

3
MOAO with EAGLE

• With the current baseline design, EAGLE will
• use 6 LGS and up to 5 NGS to map the turbulence
  above the E-ELT
• correct up to 20 x 2 diameter science fields
  anywhere within the central 5 diameter field
  using open-loop AO
• 250Hz frame rate
• E-ELT has a deformable secondary that will be
  used as a closed-loop woofer (GLAO-like DM)
• EAGLE is both a closed and open-loop system
• 30 ensquared energy with 75mas (H-band) required
  performance

4
MOAO with EAGLE big questions

• Can we achieve tomographic reconstruction to the
  required accuracy over such wide fields?
• Can we reliably control a DM in open-loop?
• How do we calibrate the system?
• How accurately do we need to measure the Cn2
  profile to optimise performance?
• What is the impact of running the system with
  both open and closed loop DMs?
• How do we compensate for LGS specific effects
  that can impact MOAO performance?
• What are the principle performance drivers
  required when designing an MOAO system?
• What is the best way to combine both NGS and LGS
  WFS signals to measure tomography?
• Answer as many of these questions as possible as
  soon as possible to feed into the EAGLE design
• Some can be (and have been) answered in
  simulation or using a lab system such as SESAME

5
CANARY concept
6
CANARY Aims

• Perform NGS then LGS based tomographic WFSing
• Perform open-loop AO correction on-sky
• Develop calibration and alignment techniques
• Fully characterise system and subsystem
  performance
• Create a single MOAO channel EAGLE as closely as
  possibly using the 4.2m William Herschel
  Telescope
• Effectively a 1/10th scale model of E-ELT using a
  10km Rayleigh LGS

7
CANARY phased development

• Based around a set of reconfigurable optical
  modules to allow easy changes between three
  CANARY phases
• Phase A Low-order NGS-only MOAO (2010)
• Phase B Low-order LGS MOAO (2011)
• Phase C High-order LGS NGS MOAO (2012)
• All phases will include an extensive calibration
  and diagnostics package

8
Diagnostics and Performance monitoring

• On-axis NGS WFS behind AO corrected focal plane
  (Truth Sensor)
• On-axis NIR imaging camera (Science Verification
  Camera)
• High-order high-bandwidth DM figure sensor
• SLODAR analysis performed using open-loop WFSs
• External turbulence profilers
• SLODAR
• MASS-DIMM
• Telescope simulator
• Turbulent phase screens
• NGS and LGS alignment and calibration sources

9
Phase A NGS MOAO

• Components
• Low-order 8x8 DM
• 3 x L3CCD open-loop NGS WFSs
• Open-loop optimised Fast Steering Mirror
• Hardware accelerated Real Time control system
• NGS MOAO Calibration Unit

10" Truth sensor IR camera FOV
NGS WFS
NGS WFS
NGS WFS
2.5 Derotated WHT field
10
Phase B Low-order LGS MOAO
3 x NGS WFS
Figure Sensor
GLAS Laser
LGS Rotator
GLAS BLT
Diffractive Optic
NGS Pickoffs
NGS FSM
Low-order DM
LGS Dichroic
WHT Nasmyth
GHRIL Derotator
LGS Pickoffs
Science Verification
Truth Sensor
Calibration Unit
1.5 Diameter LGS asterism
LGS FSM
Phase B Low-order LGS MOAO
4 x LGS WFS

• New modules include
• Electronically shuttered LGS WFS CCD
• Modified GLAS launch system
• LGS dichroic and relay system
• LGS MOAO Calibration Unit

LGS WFS
11
Phase C High-order LGS MOAO

• Closest resemblance to proposed EAGLE MOAO
  implementation
• Largest upgrade here is to the RTCS. From Phase B
  we have
• 2 times increase in pixel bandwidth
• 5 times increase in slope bandwidth
• 17 times increase in actuator bandwidth

12
Optomechanical design
13
Phase A optical design
Output focal plane
Truth Sensor focal plane
Input Focal Plane
Science Verification Camera focal plane
14
Phase B optical design
LGS TT mirror
NGS WFS placed at corrected focal plane
Acquisition camera moved to input focal plane
15
Phase C optical design concept
LGS WFS(s) moved behind closed-loop DM
Possible locations of MEMS MOAO DM
16
NGS WFS Assembly
17
Telescope Simulator
18
Subsystem performance
19
Open-loop DM Control

• 4 open-loop error with hard PZT DM demonstrated
  in laboratory with SESAME
• 40nm RMS error if a 1000nm RMS DM surface is
  requested
• Figure sensor could be used to control any long
  term drifts in DM surface shape
• Will introduce some additional latency
• Has been used with a Xinetics DM and produces a
  similar surface error to the hard PZT DM
• Open loop control of a DM doesnt seem to be a
  problem for CANARY low-order DM
• High-order MEMS DM open-loop control has already
  been demonstrated

20
Subsystem performance LGS Launch

• Test system installed on WHT and tested in May
• Uses DOE in GLAS launch system to create a 4 star
  asterism (MMT approach)
• Several possible asterisms available by changing
  DOE
• 10 to 90 diameter asterisms (takes about 15
  minutes)
• 80 of light into 4 diffracted LGS beams but
  altitude is lowered c.f. GLAS
• Still want an upgraded laser to increase WFS SNR
• Software problem with LGS detector meant range
  gated images couldnt be obtained

Non-gated image of 40 LGS radius asterism at
6.7km
DOE mounted in rotation stage at GLAS BLT
entrance
21
RTCS

• Hybrid FPGA-CPU Realtime Control System
• FPGA pixel processing developed for HOT and
  SPARTA
• Reconstructor in CPU
• DM control in CPU
• Currently runs at Phase A/B at 300-400Hz using a
  single threaded reconstructor pipeline
• Latency and jitter to be measured
• Upgrade required to cope with high-order LGS WFSs
  and DM in Phase C
• Parallelise reconstructor
• GPU acceleration

22
RTCS overview
23
System performance
24
Phase A Performance

• Monte-Carlo simulations performed using
  independent codes in Durham and Paris
• Single open-loop DM
• 8x8 actuators
• DM (and science path) on-axis
• 3 x NGS WFSs
• Off-axis (30 to 90)
• 7 x 7 subapertures
• 0.1e- read noise
• Mv 8 to 14
• 250Hz frame rate
• Representative summer La Palma turbulence profile
  used1
• r0 12cm
• 45 _at_ 0km
• 15 _at_ 2.5km
• 30 _at_ 4km
• 10 _at_ 13.5km

1 Fuensalida et al, RevMexAA, 31, 84-90 (2007)
25
Simulated Performance
26
Error terms

• Principle term is tomographic reconstruction
  error
• 30 radius means metapupils at highest turbulent
  layer are almost completely separated
• 30 is still pretty small to find a 4-star mv
  12 asterism
• Have identified several suitable targets within a
  2.5 diameter FOV observable between June-October
  
• Will be even worse with the 10km Rayleigh LGS at
  Phases B and C
• Requires the external turbulence profiling to
  determine how much of the turbulence is above the
  LGS
• The Truth Sensor will be used as the principle
  system diagnostic
• Science camera can be used when the turbulence
  cooperates
• gt60 turbulence in the ground layer is often
  observed at the WHT

27
System Calibration
28
Phase A calibration

• Interaction matrix measurement using a reverse
  path calibration source
• On-axis point source pointing backwards at output
  focal plane can be observed by each NGS WFS in
  turn
• Requires stable pupil image at lenslet array
  across full FOV
• Or use TS to measure DM influence functions
• Observe ground-layer only turbulent sources
  within the telescope simulator with NGS WFSs and
  TS
• Translate TS measure influence functions to each
  DM
• Or measure matrices on-sky
• Learn and Apply method from Fabrice Vidal first
  thing this morning

NGS WFS pickoff prism
From reverse path calibration source
To WFS
From telescope
29
Other calibration tasks

• Field dependent aberrations
• Pupil image stability is lt1/100th pupil diameter
• Monitoring and compensation changing field
  aberrations
• Non-common path error compensation
• Deployable point sources in most focal planes
• Some pointing backwards for reverse path
  calibration
• WFS linearity/gain optimisation (for WCOG etc.)
• Use sources in NGS focal plane
• NGS pickoff positioning accuracy
• Confirm with full field acquisition camera
• Detector calibration
• At Phase B/C
• LGS WFS offsets/centroid gain
• Range gate setting and optimisation
• LGS WFS interaction matrix
• To be developed further during the Integration
  and Testing phase
• Runs from October 09 April 10

30
Conclusions

• Already answered some of the big questions that
  MOAO with EAGLE raises
• Open-loop DM control
• Several calibration schemes proposed
• CANARY will have the capability to answer the
  remaining ones by demonstrating and testing
  wide-field LGS tomographic AO
• Critical subsystems are being testing and the
  initial integration phase is about to begin
• Were on track to go on-sky mid 2010 with the
  Phase A NGS tomography experiment
• Phase B design to be reviewed at the end of this
  year

31
The CANARY team
32
CANARY capabilities

• CANARY can
• Perform, calibrate and characterise accuracy of
  open-loop LGS tomography on-sky
• Measure/monitor everything to make sure we
  understand performance of each component as well
  as the system as a whole
• Develop alignment and calibration techniques
• Combine several off-axis NGS and LGS WFSs to map
  the turbulence
• Eventually use a closed-loop woofer and open-loop
  tweeter
• Emulate arbitrary LGS intensity profiles and
  elongation
• CANARY cannot
• Reach EAGLE performance goal
• Match the total number of subapertures/actuators
  within EAGLE
• Match the exactly LGS/NGS FOV afforded by the
  E-ELT
• Take advantage of the multiplex normally afforded
  by MOAO only a single channel