Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://site2010.sai.msu.ru/static/doc/GHerriot_site2010.ppt.pdf Äàòà èçìåíåíèÿ: Mon Oct 18 13:30:52 2010 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 20:10:28 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: europa

Using Site testing data for Adaptive Optics simulations Kislovodsk, October 2010
1Glen

Herriot, 1David Andersen, 1Jean-Pierre VÈran, 2Brent Ellerbroek, 2Luc Gilles, 2Lianqi Wang 1National Research Council Canada ­ Herzberg Institute of Astrophysics 2TMT Project Office, Pasadena

TMT.AOS.PRE.10.074.REL01

1


Outline
TMT / NFIRAOS Site Testing Parameters and their value for Adaptive Optics Simulations Sky coverage
­ Performance models vs season, site

DM Stroke requirement Diameter of Laser launch telescope Sodium layer structure
­ Matched filters ­ Meteor tracking

AR model of seeing
­ Centroid gain estimate in real time
TMT.AOS.PRE.10.074.REL01 2


NFIRAOS on TMT Nasmyth platform

Instruments

Space envelope Allocation for NFIRAOS

Current Design
3

TMT.AOS.PRE.10.074.REL01


Working at the Diffraction Limit Thirty Meter Telescope

TMT.AOS.PRE.10.07x.DRF01 .0x 4.REL01

4


NFIRAOS Top-Level Requirements
Throughput 85%, 0.8 to 2.5 mm Background Thermal emission < 15 % of sky and telescope Wavefront Error 187 nm RMS on-axis, and 191 nm on a 10" FoV Sky coverage 50 per cent at the Galactic pole
Differential photometry 2% for a 2 minute exposure on a 30" FoV at = 1 m Differential Astrometry 50 mas for a 100 s exposure on a 30" FoV in the H band Available from standby <10 minutes Acquire a new field < 5 minutes TMT.AOS 1 per cent 5 Downtime unscheduled <.PRE.10.074.REL01


NFIRAOS Architecture
Atmospheric tomography with six laser guide stars Near infra-red tip/tilt & focus sensing on 3 "sharpened" natural guide star images, within client instruments Multi-conjugate wavefront correction (also helps sky coverage) Minimum surface count (7 reflections + B/S + window) System cooled to -30 Celsius
TMT.AOS.PRE.10.074.REL01 6


Fundamental Design Parameters NFIRAOS
2 arcminute field 6 Laser WFSs order 60x60 in a 70-arcsecond diameter asterism
­ Polar Coordinate CCDs ­ 204792 pixels 5792 gradients per WFS

Control also uses client instruments' Wavefront sensors: 1 Tip/Tilt/Focus and 2 Tip/Tilt
­ sensing near-Infrared natural guide stars at 10 - 800 Hz.

Two Piezo Stack DMs of 63x63 and 76x76 actuators
­ DM0, optically conjugate to ground, on Tip/Tilt stage ­ DM11, conjugate to 11.2 km.

Real Time Controller solves 35K LGS WFS slopes x 7000 DM actuator tomography problem at 800 Hz.
TMT.AOS.PRE.10.074.REL01 7


NFIRAOS Opto-mechanical Layout
Input from Telescope OAP OAP OAP OAP Output to science instruments and IR T/T/F WFSs OAP Beamsplitter 2 Truth NGS WFSs 1 60x60 NGS-mode WFS LGS Trombone 76x76 DM at h=11.2 km Turbulence Simulator Phase Screen LGS Source simulator Science light Visible light Laser light

63x63 DM at h=0 km on tip/tilt stage 6 60x60 LGS WFSs

TMT.AOS.PRE.10.074.REL01

8


Parameters of interest for Adaptive Optics
r 0 Seeing and evolution of seeing vs. time 0 ... n Isoplanatic Angle, generalized for N DMs L 0 Outer scale of turbulence 0 time constant for turbulence evolution Cn2 vs altitude
­ and time evolution of Layers' strength vs time

Wind speed vs altitude Ground Level Wind-speed ­ windshake vs dome seeing Sodium layer structure, abundance and time variation Ground level temperature and variation with time Sky transparency vs time.
TMT.AOS.PRE.10.074.REL01 9


What is the interest of Adaptive Optics in r 0 Seeing ?
r
­ number of actuators needed on DMs And number of subapertures on Wavefront sensors ­ Stroke on actuators ­ Laser guide star power required ­ Sky coverage (probability of achieving astronomy) ­ Computing power in real time computer
0

Seeing Affects

Time evolution of r 0 affects update rate and accuracy of background tasks to optimize Adaptive optics control loops.
TMT.AOS.PRE.10.074.REL01 10


Value for Adaptive Optics in L 0 Outer scale of turbulence?
L 0 Outer scale of turbulence Affects DM stroke required Affects Tip/Tilt Focus stroke and bandwidth
­ Smaller L0 means less stroke needed for the same r0. ­ Smaller L0 means less energy in low modes and low frequencies ­ for both optical and numerical simulations

Affects Phase screens for turbulence simulation

Time evolution of L0 affects background tasks, which optimize Adaptive optics control loops.
TMT.AOS.PRE.10.074.REL01 11




0,,

n Isoplanatic Angle generalized for N DMs

0 , , n Isoplanatic Angle, generalized for N DMs Affects corrected field of view Thus affects sky coverage
­ Because tip/tilt/focus stars should be found in corrected field.

Affects ­ And Affects ­ And

optimal number of DMs their ideal altitude of conjugation number of Laser Guide Stars their spacing on the sky

Affects number and location of optical phase screens in turbulence simulator
TMT.AOS.PRE.10.074.REL01 12


0 time constant for turbulence evolution
­ Readout rate of WFS ­ Laser power, read noise of WFS ­ Computer speed of real time controller

0 affects bandwidth for AO control system

TMT.AOS.PRE.10.074.REL01

13


Parameters of interest Cn2 vs altitude
Cn2 vs altitude
­ Determines Number of layers in tomographic reconstruction and thus computing power ­ Defines DM quantity and Optimal altitude of conjugation ­ Good initial data allows quick settling of tomography algorithm to final value to begin science exposure ­ Determine potential effectiveness of a Ground Layer AO system.

TMT.AOS.PRE.10.074.REL01

14


Parameters of interest Wind speed vs altitude
Wind speed vs altitude
­ Frozen flow ­ Predictive filter methods are desirable, ­ But how effective are they? Simulations can tell us, providing that we have good data.

TMT.AOS.PRE.10.074.REL01

15


Parameters of interest Wind speed vs altitude
Ground Level Wind speed
­ Windspeed data feeds Dome Computational Fluid Dynamic wind force models, which are applied to TMT structural finite element models and controls model of telescope and mirror segments. ­ Resulting windshake is disturbance input to NFIRAOS simulations of performance and sky coverage

Dome computational fluid dynamics and heat transfer models create dome seeing voxel (volume elements) maps within dome. Ray tracing through dome voxel dome creates phase screens
­ Input to Adaptive Optics simulations.
TMT.AOS.PRE.10.074.REL01 16


Parameters of interest Ground level temperature vs time.
Ground level temperature variation with time
­ Temperature variation of telescope and dome cause dome seeing ­ Near-IR background flux from warm telescope optics increases integration time for background limited objects. ­ Point Source sensitivity calculations affected

TMT.AOS.PRE.10.074.REL01

18


Fraction of nights with Sub-visible cirrus causing Fratricide and Scattering
Four scattering effects studied: Rayleigh, ozone, aerosol, cirrus
­ Rayleigh scattering induces fratricide between LGS WFSs for Central Launch
0.8

I P ( I nf ) e N (0,1) P ( I

bkg

) cal I

bkg

­ Real-time updates at ~0.1Hz are expected to provide required calibration accuracy to better than 80%

Incremental WFE (nm) Zenith angle (deg) 0 30 % affected subaps 0.4 0.7 0% calib. 12 20 80% calib. 1 5

Computed by integrating the atmospheric backscattered light intensity profile (volume scattering coefficient) over altitude, subaperture area, and pixel FoV. 45 1.5 39 10 60 4.6 117 31

Ozone, aerosol and cirrusTMT.AOS.PRE.10.074.REL0ntary signal level variations contribute to mome1 19 for both CL and SL: ~23 nm RMS for 20% reduction


Telemetry from AO systems continues to "survey" site.
Telemetry from Adaptive Optics Systems can continue to monitor sites. Classic AO System
­ Gemini Altair outputs r0 and L0 based on Telemetry

­for Gemini Gpi AO system ­ Poyneer & Veran ­
­ Simulations using Gemini Altair and NICI Telemetry says GPi can determine Number of atmospheric layers and wind speed for each ­ But not the altitude and strength of each layer

While there is a good fraction of turbulence that appears to be frozen flow, there is also a significant portion that is not. All proposed AO predictive control schemes currently assume frozen flow...
TMT.AOS.PRE.10.074.REL01 20


Real-Time Cn2 Profile Estimation for Optimal Tomographic Wavefront Reconstruction
SLODAR-like method correlates pseudo open-loop measurements from a pair of the 6 NFIRAOS LGS WFSs Eliminates sensitivity to LGS tip/tilt/focus by using second-order differences of gradients Computationally efficient and convergent in a few hundred frames at 800Hz Vertical resolution h1 2~ km ~ | | 1 / . 5





·6 layers estimated from 11 baselines ·Solves linear system of the form
5 A , x/3 x b k rk 0, A computed using Fourier

technique
21

TMT.AOS.PRE.10.074.REL01


TMT Error Budgeting and Performance Analysis
Comprehensive evaluation of TMT AO architecture
­ Wavefront disturbances due to atmosphere/telescope/NFIRAOS/ instruments ­ NFIRAOS wavefront sensing and correcting hardware ­ LGSF and OIWFS components ­ NFIRAOS processing algorithms

Performance evaluation as a function of seeing, zenith angle, field of view and galactic latitude Estimates developed through a combination of:
­ ­ ­ ­ Integrated AO simulations Side analyses Budget allocations Lab and LIDAR experiments .10 TMT.AOS.PRE

.074.REL01

22


Simulation Tools for LGS Performance Analysis and Sky Coverage Evaluation
LGSs Science Object Wavefront Correctors LGS WFS (6) LGS Perf. Eval. NGS Mode Fitting Complex Image Field Asterism Selection OIWFS (3)
500 guide star fields

7 x 7 NGS Array (Asterism Pool)

POL LGS grads (~35K) TT/DF Removal LGS Tomography DM Fitting (~7K actu.) NGS Mode Removal

+ -

NGS Recons.
NGS Mode WFE CDF

Physical optics model Type II Woofer/Tweeter control Telescope windshake PSD

Sky coverage Post-processor

TMT.AOS.PRE.10.074.REL01

23


Key Results Over the Last Two Years
Performance analysis for Mauna Kea confirms that performance requirements are met:
­ 187 nm on-axis at zenith with median seeing and 50% sky coverage at the Galactic Pole met with 83 nm RMS margin in quadrature ­ Based upon detailed time domain simulations of NFIRAOS, including WFSs, DMs, RTC, and telescope models

Sky coverage has been evaluated and optimized in detail:
­ ­ ­ ­ Physical optics modeling of OIWFSs Monte Carlo simulations over 500 guide star fields Evaluation as a function of zenith angle and seeing OIWFS Pixel processing and temporal filtering algorithms studied in detail
TMT.AOS.PRE.10.074.REL01 24


Turbulence Parameters for 25% & 50% Mauna Kea conditions
Altitude (km) Wind Speed (m/s) 0 5.6 0.5 5.8 1 6.2 2 7.6 4 13 8 19 16 12

MK13N 25% profile, r0= 27.4 cm, 0 =2.7", fG=15.9Hz Weights (%) 32 15 4.7 4.1 16 11 18

MK13N 25% profile, r0= 19.9 cm, 0 =2.2", fG=21.7Hz Weights (%) 29 18 6.6 7.8 14 12 13

TMT.AOS.PRE.10.074.REL01

25


Winds aloft, and Cn2 for Median and Good Seeing at Mauna Kea

Left: Turbulence profile relative weights (50% blue, 25% red). Right: Wind speed profile used in conjunction with the turbulence profiles shown on the left. from TMT.AOS.TEC.10.009.DRF01

TMT.AOS.PRE.10.074.REL01

26


NFIRAOS PSF for Mauna Kea

TMT.AOS.PRE.10.074.REL01

27


TMT NFIRAOS feeding multi-slit spectrograph (IRMS)
Estimated ensquared energy curves 50% Mauna Kea turbulence conditions

TMT.AOS.PRE.10.074.REL01

28


Sky Coverage Analysis
Performance characteristics of H2RG OIWFS detector modeled in detail Matched filter pixel processing algorithms and type II woofer-tweeter control law have been tuned to optimize performance Requirements met with margin at zenith Off-zenith performance limited by physical optics effects
­ ­ Lower NGS Strehls, smaller 0 and 2, no diffraction-limited PSF core at large offsets Unobserved previously with geometrical OIWFS models excluding physical optics effects

TMT.AOS.PRE.10.074.REL01

29


Performance vs seasons
Turbulence spatio-temporal parameters versus seasons starting with winter (Dec.- Feb.), for Mauna Kea (M1-M4) and Cerro Armazones (A1-A4). At Zenith and = 500nm

TMT.AOS.PRE.10.074.REL01

30


RMS WFE (nm) versus zenith angle Mauna Kea and Cerro Armazones.
Black red blue green curves correspond respectively to the winter/spring/summer/fall seasons

Mauna Kea
TMT.AOS.PRE.10.074.REL01

C. Armazones
31


DM stroke requirements
Histogram of the DM actuator commands OPDs of the ground and upper DMs for a variety of turbulence profiles that have similar 90th percentile 0 But quite different values of r0, ranging from 0.07 m to 0.193 m. The outer scale is 30 m. The upper DM has more or less similar command distributions for all of the profiles The ground-conjugate DM has broader histograms for smaller values of r0.
TMT.AOS.PRE.10.074.REL01 32


Deformable Mirror Stroke Requirement
Histograms of actuator commands

TMT.AOS.PRE.10.074.REL01

33


Wavefront Error vs DM stroke for Classic AO (single DM system)
L0 = {30, 60} m and r0 = {0.07, 0.1, 0.13, 0.15} m

r0 L0
If L0 is large for a given r0, then DM requires more stroke to achieve the same wavefront same error

TMT.AOS.PRE.10.074.REL01

34


Site Survey Temperature Data
Site survey data of mountain-top temperature drives AO system temperature for low background observations. Median Temperature on Mauna Kea is 2.3 C Requirement of NFIROAS adding < 15% of sky and telescope background in K band implies cooling NFIRAOS.

TMT.AOS.PRE.10.074.REL01

35


Temperature vs Emissivity
·Observing time decreases directly with decrease in thermal background ·Cooling NFIRAOS cuts observing time by a factor of 2.4 in K band Temp.
Background vs Wavelength Flux Allowable Temperature vs Emissivity

15%(Telescope + Sky) K Band

Just meet Specification

-26 C -30 C
NFIRAOS Design
TMT.AOS.PRE.10.074.REL01 36

2.2

2.5 Wavelength µm

18%

Emissivity


Turbulence Simulator
Phase screen deployed into science path Eliminates separate turbulance simulator in front of window We are investigating MRF polishing of the phase screens ~ 360 x 750 mm Turbulence also added to DM commands Reproduces r0 & 2
Window LGS sources

Phase Screen

TMT.AOS.PRE.10.074.REL01

37


Turbulence Simulator screen
Optimal altitude & strength of screen to build into AO system.
­ Estimated by simulations based on site survey data.

Candidate Altitude

TMT.AOS.PRE.10.074.REL01

38


Sodium Density Profiles from UBC Vancouver Lidar

TMT.AOS.PRE.10.074.REL01

39


Power Spectrum of Sodium Altitude from UBC Lidar -

TMT.AOS.PRE.10.074.REL01

40


Na Layer Range Tracking
Error in Na layer range is tracked by the OIWFS
­ 4 nm / meter of error in Na range estimation

But OIWFS sampling frequency can be low (median 90Hz), so errors will occur due to delay Error budget updated via latest UBC Lidar measurements
At 90 Hz OIWFS sampling rate the residual defocus error is 11.8 nm rms m2/Hz Sodium altitude power spectrum

nm RMS

10 5

Prior art

Residual focus WFE vs. OIWFS sample rate 0 100 Hz 400

UBC Lidar

TMT.AOS.PRE.10.074.REL01 0.001

Hz

10 41


Meteor Trails

TMT.AOS.PRE.10.074.REL01

42


Simulation results from Sodium data
Sodium movies played into simulations, in computer and on UVic AO lab bench to assess:
­ Residual errors from meteor transients. ­ Power consumption of focusing trombone 60 W during meteor transient (early result to be confirmed) ­ Determine suitable update interval for background tasks, and residual errors from sodium variability

TMT.AOS.PRE.10.074.REL01

43


Adaptive Vibration Compensation Algorithm
Efficiently compensates for the effects of vibrations using a local oscillator locked in phase, amplitude and frequency that injects a counter vibration on TTS and tracks changing conditions.
Tip/Tilt residual (mas rms)

WFS freq 800 Hz Control Type I control Type II control
Type II + AVCA

90 Hz 23.06 21.30 15.51

40 Hz 14.29 14.30 14.30

8.210 8.810

Input Tip/Tilt disturbance: Atmosphere: r0=15cm, L0=30m Windshake: 50%, rms=7.5mas Total: 18.8mas rms 29.5Hz vibration: 13.3mas rms Total disturbance: 23mas rms

Type II + Notch 2.944

0.00434 0.0919 0.303 Offers superior performance and works even at low sampling frequencies of OIWFS (TT WFS) Performance is only reduced when WFS sampling frequency ~ vibration frequency due to aliasing
TMT.AOS.PRE.10.074.REL01 44


Time Variability of r0
Autocorrelation of log(r0) Power spectrum of log(r0)

TMT.AOS.PRE.10.074.REL01

600 min

45


r0 time series ­ autoregressive model built from autocorrelation of r0

Bad seeing case

1 hour
·Avoids having to choose a "representative" night time series. ·Time series used in simulations of ·NGS-mode WFS centroid gain estimator (background task) · image smearing during long exposures to assess astrometry accuracy.
TMT.AOS.PRE.10.074.REL01 46


Desirable to have autoregressive model of the evolution of layers' strength
Layers' strength vs time
­ Would like to assess importance of good initial guess of layer strength for tomography, ­ Would like to estimate update rate needed for background tasks

However, the technique for r0 just described does not work for individual layers of TMT site data.
­ too noisy per-layer TMT data.. negative numbers sometimes.

TMT.AOS.PRE.10.074.REL01

47


Laser Launch Telescope Location
End to end Monte Carlo physical optics simulations
­ Side launch provides ~20 nm better Wavefront error, but at increased cost and complexity. ­ 4 laser launch telescope (LLT) configurations investigated. ­ Circles indicate the associated guide star (GS) asterism. Each GS is projected by the closest LLT, in all cases. TMT Baseline

Incremental WFE w.r.t baseline

TMT.AOS.PRE.10.074.REL01

48


Laser Launch Telescope Diameter
LLT diameters 0.1, 0.3, 0.4, 0.5 and 0.7 m, r0 0.10, 0.15, 0.20 m, { 75%, 50%, 25% } seeing, LGS signal levels of 250, 500, and 1000 photons detected /subaperture/frame at 800Hz, Nominal sodium profile Nominal Cn^2 profile for Mauna Kea
Incremental Wavefront error vs Launch telescope Diameter

TMT.AOS.PRE.10.074.REL01

49


Acknowledgements
The TMT Project gratefully acknowledges the support of the TMT partner institutions. They are the Association of Canadian Universities for Research in Astronomy (ACURA), the California Institute of Technology and the University of California. This work was supported as well by the Gordon and Betty Moore Foundation, the Canada Foundation for Innovation, the Ontario Ministry of Research and Innovation, the National Research Council of Canada, the Natural Sciences and Engineering Research Council of Canada, the British Columbia Knowledge Development Fund, the Association of Universities for Research in Astronomy (AURA) and the U.S. National Science Foundation.
TMT.AOS.PRE.10.074.REL01 50