telescope wavefront errors · the scaling between wavefront slopes and wfe is shown in table 3. wfe...

TelescopeWavefrontErrorsHenriBonnet

1E-ELTDataSimula;onWorkshop,2016-04-15

TasksofWFCatE-ELT

•   HelpSystemEngineeringdevelopandmaintainthetechnicalbudgets

•   DevelopControlStrategy•   DefineWFCI/Ftoinstruments.


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.


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,…)


DifferencesVLTvs.E-ELT

•   ThewavefrontdeliveredbytheVLTisseeinglimited:–  Wavefronterrorscreatedbythetelescopearecon;nuousandslow

–   Alwaysaminordistor;ontothepowerspectrumofthefreeatmosphere.

–   Outstandingexcep;on=vibra;ons•   10to15masrmsof;p-;ltatafewharmonicfrequencies.

•   VLThasfewsensorsandactuators–   Failureofanequipment=down;me

•   E-ELTisnotlikethis.


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.


WindShake:M1


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.


PlateScale/fieldrota;on

•   ThegoalisthatCCSholdstheplatescalefor1hour.

•   Fieldrota;onmaynotbepredictableatthediffrac;onlimit.

=>Instrumentsmayneedtoincorporatesecondaryguidingforplatescale/fieldrota;on.


M1


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


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.


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?


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.


Opera;onalincidents

•   ES/PACTfailures•   Missingsegments


telescope wavefront errors · the scaling between wavefront slopes and wfe is shown in table 3. wfe...

Documents