telescope wavefront errors · the scaling between wavefront slopes and wfe is shown in table 3. wfe...
TRANSCRIPT
TelescopeWavefrontErrorsHenriBonnet
1E-ELTDataSimula;onWorkshop,2016-04-15
TasksofWFCatE-ELT
• HelpSystemEngineeringdevelopandmaintainthetechnicalbudgets
• DevelopControlStrategy• DefineWFCI/Ftoinstruments.
2E-ELTDataSimula;onWorkshop,2016-04-15
Howwedoit• DefineaWFCplan,describing:
– howwephaseM1– howwemaintainthetelescopecollima;on– howwerejectthedynamicperturba;ons
• Evaluateerrorpropaga;ons->Simulate– FEM– Dynamicsimula;ons– Raytracingofsegmentedmodel– AOsimula;ons
• Widerangeofspa;alandtemporal;mescales– Noendtoendsimula;ons– Simula;ontoolsarecustomizedandinterfacedtooneanotherforeachques;onaddressedbytheteam.
3E-ELTDataSimula;onWorkshop,2016-04-15
Control Algorithms
Input Wavefront
Delivered Wavefront
Optical Propagation
Atmosphere Mirrors
Manufacturing
Integration
ActuatorsMetrology
Model Errors
Noise
Structure
Adaptive Optics35Hz
ES Loop1 Hz
Main Axes Loop1 Hz
Low Order Optimization 0.1 Hz
Calibrations << 1Hz
Turbulence
Dynamic Perturbations
Wind
Heat
Vibrations
Gravity
Internal
Sky
Identify Errors
Allocate Errors
Write Requirements
Design
Model
Evaluate Error Propagation
Define Control Strategy
PerformanceAnalysisPurpose&Process
E-ELTDataSimula;onWorkshop,2016-04-15 5
WFCProducts• Sensi;vityanalyses
– Providestheconnec;onbetweensub-systemsrequirementsanderrorbudget.
• Calibra;onandwavefrontcontrolbaseline• WFinterfaces
• Requirementstocontrolequipments– WavefrontSensors– Metrologies(e.g.EdgeSensors)– Actuators(stroke,resolu;on,…)
6E-ELTDataSimula;onWorkshop,2016-04-15
DifferencesVLTvs.E-ELT
• ThewavefrontdeliveredbytheVLTisseeinglimited:– Wavefronterrorscreatedbythetelescopearecon;nuousandslow
– Alwaysaminordistor;ontothepowerspectrumofthefreeatmosphere.
– Outstandingexcep;on=vibra;ons• 10to15masrmsof;p-;ltatafewharmonicfrequencies.
• VLThasfewsensorsandactuators– Failureofanequipment=down;me
• E-ELTisnotlikethis.
7E-ELTDataSimula;onWorkshop,2016-04-15
Windshake:Al;tudeStructure
• FromVLTtoE-ELT:– Largerstructure
• Lowereigenfrequencies• Highersensi;vitytodynamicperturba;ons
– Kbanddiffrac;onlimitfrom56to12mas.• Largeeffortinvestedduringalldesign
phases– DrivertoMainStructurerequirements– 2stagecontrolstrategy(M4+M5)with
enhancedrejec;onatlowfrequency• Resultedinsa;sfactoryperformance
– 1.6masrmsinstandardcondi;ons• E-ELTisVLT-likeinthisrespect.
8E-ELTDataSimula;onWorkshop,2016-04-15
WindShake:M1
E-ELTDataSimula;onWorkshop,2016-04-15 9
Telescopecollima;on• M2
– Highop;calsensi;vity– Largeiner;a
• Resolu;onofposi;oningsystemincompa;blewithAOperformance
• PhaseBdesignconductedunderconstrainedthatM2wouldnotbereposi;onedduringobserva;on(1hour).
• Studiesconcludedthatthiswasnotdoable
• L1requirementrelaxedto1reposi;oningevery5minutes.– Performanceandstrokebudgets
nowok– 20mas(postAO)transientatlow
orderop;miza;on.
Page : 10/25
M2 Hexapod Dynamic Requirements 10th June, 2013
Kronodoc#: ESO-214629
File: M2HexapodDynamicRequiments_v6.doc
Authors: H.Bonnet, M.Mueller, L.Pettazzi
for the active model and at ~9.5Hz for the passive one) which are peculiar feature of the ELT not present in the VLT DSM test bench and therefore do not reproduce exactly all the features of the response observed in Figure 4. Moreover a dead band controller has been included to avoid chattering around the target position. Such a solution is a standard measure against friction induced perturbations such as hunting (see RD5)
Figure 6: Simulated step response of the ADS HP
This hexapod model is used with two different M2 models representing both the passive and active M2 unit solution.
6 Observation Scenario This section is about the impact to the image quality of the dynamic performance of the 2 models.
The problem is the transient perturbation caused by a repositioning of M2. We consider the scenario where this is done every 5 minutes. This time scale is a compromise between 2 needs:
- It needs to be shorter than the current 1 hour baseline, which does not work. 5 minutes is about 1/10th of an hour.
- the low order optimization of the telescope will run at the same rate as the VLT active optics (cycle duration ~30 sec). However, the transient perturbation induced by a repositioning of M2 may induce a temporary degradation of the wavefront quality (after AO correction) larger than the specified optical quality. We hope that a cycle of 5 minutes will be suitable for all applications:
Page : 18/25
M2 Hexapod Dynamic Requirements 10th June, 2013
Kronodoc#: ESO-214629
File: M2HexapodDynamicRequiments_v6.doc
Authors: H.Bonnet, M.Mueller, L.Pettazzi
Figure 12: tip-tilt (top) and high order wavefront errors (bottom) time series after the execution of the M2 corrections generated by the Low Order Optimization. The responses to all commands generated during the scenario (tracking at maximum altitude rate from zenith to 30 degrees elevation) are superposed in one plot. The results are evaluated at the science focal plane after AO. The scaling between wavefront slopes and wfe is shown in Table 3.
WFE Slopes rms mean Peak Deviation Tip-tilt 1 µm 20 mas 0 Focus 1 µm 0 77 mas Coma 1 µm 37 mas 200 mas
Table 3: relation between slopes and wfe for low order aberrations.
The maximum excursion in tip-tilt is 12mas, well within the FOV of the wavefront sensors (3 arcsec).
The high order wavefront error is also negligible (3 mas peak value, assuming that the peak at 15nm is pure coma) in comparison to the FOV of the WFS.
Conclusion: this hexapod model is compatible with the scenario of one M2 correction every 5 minutes, if the AO runs continuously during the correction.
For information the tip-tilt PSD after AO computed over the first 18 seconds after the command was sent are shown in Figure 13.
10E-ELTDataSimula;onWorkshop,2016-04-15
PlateScale/fieldrota;on
• ThegoalisthatCCSholdstheplatescalefor1hour.
• Fieldrota;onmaynotbepredictableatthediffrac;onlimit.
=>Instrumentsmayneedtoincorporatesecondaryguidingforplatescale/fieldrota;on.
11E-ELTDataSimula;onWorkshop,2016-04-15
M1
E-ELTDataSimula;onWorkshop,2016-04-15 12
E-ELTPhaseBFinalReview,September22nd2010 Slide
13
SegmentMisfigurePolishing
PolishingRequirements(E-SPE-ESO-300-0150.2)
WarpingHarnessModelandFiDngError
1
2
3
FirstpolishedprototypeatSAGEM
Higherorders:1/f2PSDPrototypemeasurement
Simulatedsamples
Phasing
• PhasingproceduredemonstratedatGTC.
• Baseline:– updateofphasingsolu;onevery2weeks.
– Localmetrologies(EdgeSensors)maintainthephasingbetweencalibra;ons
14E-ELTDataSimula;onWorkshop,2016-04-15
Op;calPhasing
Measured OPD (nm)
1
2
3
4
5
6
7
8
9
10
11
1213
14
1516
17
18
19
20
21
22
23
24
25
26
27
28
29
30 31
32
33 34
35
36
0
50
100
150
200EELT
GTCData
GTCExperimentscheme
þ
☐
þ
þ
þ
þ
Valida;onatGTCwithESOtechnical
;me
ErrorBudgetalloca;onØ basedon100nmOPD
measurementnoiseØ WFEcompensatedbySCAOØ LoworderoffloadedfromM4Ø ResidualWFE=35nm.
GTCperformancees;ma;on(July2010)Ø measurementnoise~40nmØ OPDresidual~40nm
E-ELT Programme Offline Machine Modus Operandi Template
Alignment Reference
Post-Processing
Sensitivity Matrices
Data
Fringes
Ray Tracing Model
Online processing Telescope
Sequencer
FTP
FTP
Guiding (manual)
Data
Scalloping• Scallopingistheresultofalarge
focuserrorinM1compensatedelsewhere(e.g.withM2).
• Consequence:mismatchbetweenradiiofcurvatureof– Segments– Segmentsassembly
• Highorderwavefronterrorwithfirstorderdiscon;nui;es.
• Scallopingbudget~35nmbutthisisconsideredatechnicalrisk.
• Riskmi;ga;on:– MakeEdgeSensorssensi;vetoM1focusmode(PSGsensors).
– GuideProbeWFScapableofobservingscallopingatpreset.
16E-ELTDataSimula;onWorkshop,2016-04-15
SegmentinplanedisplacementsPiston–Shear–GapEdgeSensors Op;calcalibra;on(Phasing)every2weeks PSGfeatures
Ø FullobservabilityofmirrorstateØ Performanceagainstgravityandthermalperturba;ons
limitedbymoun;ngerrors. Installa;onAccuracy
Ø Rota;onError=1mradTobemeasuredwithanaccuracyof0.1mrad
Ø Transla;onError=100µmDrivenbycapturerangebudget
Gravity: Zenith to 45deg
ImpactofdifferenQalgravityloadfromzenithto45deg
Impactof20Kambienttemperaturechange
amplitude
Correc;ons
ErrorOp;calPhasing
WarpingHarness SCAO
Integra;on 1mm/1mrad ✗ ✗ ✗ 4nm
Gravity fromFEA ✗ 8nm
AmbientTemperature
seasonal 6K ✗ ✗ ✗ 2nm
2weeks 4K ✗ 6nm
Temperaturegradient 1K/42m/axis ✗ 3nm
Total 11nm
Shear−Gap compensation OFF: 100 nm
Shear−Gap compensation ON: 10 nm
1mradclockingerroronsegment#429
Phasing• Difficul;es:
– LargenumberofDOF– Inplanemo;ons(PSGsensors)– Couplingbetweenin-planemo;onandESsignals.– Newsegmentseveryday(re-coa;ng)– Shapeofsegmentsbehindthespiderpoorlyobservable
– Localvibra;ons=>Locallylargediscon;nui;esinthewavefront.
Diffrac;oneffectsinWFS?
18E-ELTDataSimula;onWorkshop,2016-04-15
Spider
• Shapeofsegmentshiddenbyspiderarepoorlyobservable– Surfacediscon;nui;esattheedgesofthesesegments
– Propaga;onofphasingerror
• Spiderwidth=530mm>r0– Fragmenta;onofAOpupil
Technical Note Doc. Number: ESO-285814
Impact of spider obstructions on phasing Doc. Version: 1.1 Modified on: 2016-04-01 Prepared by:
Page: 7 of 7
Document Classification: ESO Internal [Confidential for Non-ESO Staff]
3.4 Results For each case 10 random realization were made with 5 phasing iterations for each case.
Figure 5 shows an exemplary phasing solution for each of the two cases.
Figure 5: Exemplary phasing solution for Case A (left) and Case B (right).
The table below reports the mean and the standard deviation of 10 Monte Carlo realizations with 5 phasing iterations each. These values are before AO correction.
Residual RMS WFE (mean / std)
Case A 61 nm / 29 nm
Case B 142 nm / 40 nm
4. Conclusion It shows that with an increase of the spider widths to a level where the next additional inner subapertures are vignetted the noise propagation in the phasing control results in solutions that are fragmented along the petals of M1. Consequently, there are considerable optical discontinuities between the M1 sectors that are left uncorrected.
19E-ELTDataSimula;onWorkshop,2016-04-15
Opera;onalincidents
• ES/PACTfailures• Missingsegments
20E-ELTDataSimula;onWorkshop,2016-04-15