INAF-Osservatorio di Arcetri 3-4 May 2002. Tuorla Observatory Opticon Meeting Adaptive secondary and primary mirrors for ELTs A. Riccardi 1, G. Brusa 1,2,

INAF-Osservatoriodi Arcetri

3-4 May 2002.Tuorla Observatory

Opticon Meeting

Adaptive secondary and primary mirrors for ELTs

A. Riccardi1, G. Brusa1,2 , P. Salinari1, D. Gallieni3 and R. Biasi4

1INAF-Osservatorio Astrofisico di Arcetri, Italy2CAAO, Steward Observatory, AZ, USA

3ADS International Srl, Italy4Microgate Srl, Italy



Opticon Meeting

Corrector specs for ELTs

actuators for the ground layer:

D=50m about 100,000 D=100m about 400,000

Most challenging target:Multi-conjugate adaptive correction in V band

seeing: 0.65arcsec (r0=0.16cm at 0.55mm)r01=22cm (60%) r02=28cm (40%)

SC Fitting error=0.3 (d/r0)5/3 rad2

MC Fitting error=0.3 [(d1/r01)5/3+ (d2/r02)5/3] rad2

MC Fitting error == SC Fitting error with d= r0

(d1/r01)5/3= (d2/r02)5/3 = 1/2

dk = r0k / 23/5 = 0.66 r0k d1 = 14cm, d2=18cm



Opticon Meeting

Corrector specs for ELTs

So many actuators need:

• Low cost per actuator

• Reliable device

• Easy maintenance



Opticon Meeting

Currently avail. correctors

Piezo technology

Largest unit: Xinetics, 349 act., 7mm spacing (M=20). Position actuators (failure is a major problem), limited stroke for low spatial order correction (wind effects)MEMS technologyLarge number of acts., reduced stroke, 0.3mm spacing (M=466)Field=6arcmin, M=466: Relay optics F#=F/1.2Adaptive secondary technology

MMT unit: 336 act, 30mm spacing (first light end of June 2002)LBT units: 672 act, 30mm spacing (first light mid-2004)Force (electromagnetic) actuators, Large stroke (Chopping)

See Brusa et al. Backaskog Workshop for a general theory

N act > 300



Opticon Meeting

People (some of them)

Osservatorio di Arcetri

INAF-OAA (Italy): Concept development.Testing, calibration and diagnostic software

ADS (Italy): mechanical engineering,drawing and assembly

Microgate (Italy): electronics and DSP software

Media Lario (Italy): mechanicaldrawing and assembly

CAAO-SO (USA): MMT AO

Mirror Lab-SO (USA): optics andoptical test facility



Opticon Meeting

• mirror diameter 642 mm• mirror thickness 2 mm• membrane in-plane restraint• 336 moving magnet actuators• nominal air gap ~ 40 m• reference body 50 mm thick• AL cold plate: actuators support & cooling (7 cooling channels)• support frame & interface to Hexapod• electronics cooled crates• 1Gbit/s communication link (optical fiber)• hub interfaces (power, signal & cooling)

System Layout



Opticon Meeting

MMT336 ASPHERIC SHELL

Assembled unit

Magnets

642mm diam.2mm thick

(12mm diam)

Cap. sensor armatures (ref.plate)



Opticon Meeting

• 3 crates

• 14 control boards each crate

• 8 channels controlled by each board (4DSP)

• capacitive sensor signal conditioning on the actuators (close to variable gap capacitor)

Electronics – control system



Opticon Meeting

Step response

Only Prop. - Gain=0.2N/m (40m gap) Settling time 1.5ms

CommandPosition



Opticon Meeting

Step response

FF Force

Ctrl+FF Force



Opticon Meeting

Step response



Opticon Meeting

Step response



Opticon Meeting

Turbulence compensation

nm nm

Input turbulence Residual of first 200 modeCorrection (scale X10)

r0=15cm @ =550nm, v=6m/s, 8sec

2.3m rms140nm rms on acts.Max 0.4N rms(|T|<1.2 C) WCcondterm

WNeff

/5.3..

/5.0

(Cap. sensor readings)



Opticon Meeting

Modal analysis

(Cap. sensor readings)



Opticon Meeting

Optical tests at Mirror Lab

Shimulator (Steward Obs-Mirror Lab)

Turbulence simulator (Steward)

Test of the entire AO system:Adaptive secondary+WF sensor+WF computerclosing the optical loop



Opticon Meeting

MMT336 Flattening

The flattening of the shell applying 150 mirror modes: 34nm RMSThe central rings are artifacts due to the olographic technique used for testing convex mirrors



Opticon Meeting

Closed optical loop

Open loop Closed loop @ 550Hz

MMT336+SH 12x12+WF computer

Data from G. Brusa (CAAO-Mirror Lab)



Opticon Meeting

LBT672 summary

672 actuators in 14 rings

31mm act-to-act ave. distance

642mm

911mm

MMT336

LBT672



Opticon Meeting

LBT672 electronics3x2 crates

14 ctrl board per crate (84 tot)

8 320Mflop/s 32bit-FP DSP per board (ADSP-21160)

2 actuators per DSP (336 tot.)

Total computational power:

107 Gflop/s

WF reconstruction on-board LBT672

((32x32x2)x672)x2 op in 0.5ms = 6Gflop/s (5% res.)

(8.4s to compute the FF matrix)



Opticon Meeting

Actuator spacing limit

30mm (possibly 25mm)

Capacitive sensors:3nm resolution, 100m range (DR propto area)

Actuator Efficiency: proportional to sqrt of magnet and coil volume

Interaction among magnets



Opticon Meeting

Adaptive sec. location

OWL

Euro50

Ground layer:4m M2 (31cm)

10km layer:1.5m M2 (83cm)

Ground layer:2.35m M6 (106cm)

8km layer:4.2m M5 (60cm)

K band

J-H band

I-J band

H-K band



Opticon Meeting

Actuator spacing limit

Adaptive secondary technology is available and ready for NIR correctors on ELTs



Opticon Meeting

Primary adaptive mirrorNo more problem of actuator spacing (M=1)

Economical advantages:Less expensive magnet technology

Euro50 segments: 70mm thicknessOWL segments: 80-100mm thicknessGravity load is supported in more points:Primary adaptive mirror: 10mm thickness 86%-90% less glass!!

High spatial frequency control: larger polishing residualerror can be accepted at low and mid-spatial frequencies

No Zerodur, borosilicate can be used. Cheaper!!



Opticon Meeting

2m segment layout

10cm/act

397 act/segment

Local glass stiffness: 15N/m

(LBT: 672 actuators)



Opticon Meeting

Main components

Glass:77.5kg/segHoneycomb:100kg/seg400 Actuators:80kg/seg

Total:258kg/seg

Euro50 (70mm thick.): 543kg/seg.52% larger mass!!

OWL (80mm thick.): 621kg/seg. 58% larger mass!!



Opticon Meeting

Support and maintenance



Opticon Meeting

Backplate FEA

Deformation of the honeycombwhen pushing: 238N/m(glass: 15N/m)

Gravity load: 15m PtV



Opticon Meeting

Actuator details

Axial and tangential load: astatic levers?



Opticon Meeting

Actuator prototype

< 2ms settling time



Opticon Meeting

Control electronics

Control electronics,one box per segment:smaller then LBT672

In a multi-resolutionreconstruction approachcontrol electronics canwork as WF reconstructor



Opticon Meeting

Primary mirror prototype

1200mm diam.7x340mm segments19 acts/segment



Opticon Meeting

Conclusions

• Adaptive secondary mirror technology is a proven technology that is available and ready for AO correction for ELTs in the NIR

• Effective correction in V band using non-primary mirrors requires a step up in the number of actuators with the right actuator density that is not available right now

• Adaptive secondary concept and technology can be applied to primary mirrors to achieve effective correction in V band

• Warning: delaying a solution for an effective correction in the V band for ELTs could avoid a future upgrade in this direction. Let’s think about a compatible V solution now!



Opticon Meeting

Cost goals



Opticon Meeting

Resonances in control bw

First 270 modes have res< 1kHz + local controlPhase lag of 180deg for >res if low damping => unstable

P30 “astigmatism” mode TF

37m gap

115m gap

37m gap

115m gap

High damping (18Ns/m => 40m gap) + local control PD

increasingDamping:larger PM



Opticon Meeting

Control loop

Commands from WFC

kc

DsG -

+

DM

k-th Curr. Driver + coil

CapsensLinearization

@625Hz

DSP of k-th channel

-3dB@56kHz

40kHz local loop

DSP

ii

ki cK

336

133621, cccc k

+

Feed-forward

++



Opticon Meeting

Electronic damping

100m gapno electronic damping

Electronic dampingtest on P36

G Plant Sens

D

+-c

ff

+++

PV: 0.5ms settling time

50Ns/m



Opticon Meeting

Electronic damping

Natural vs. electronic damping:• Natural damping reduces stroke• Electronic damping is limited by delay and capsens noise

New capsens for LBT:From 40kHz to 500kHz•Reduced aliasing error•Analog derivative (reduced delay)•Electronic damp. 4kHz->50kHz

G Plant Sens

D

+-c

ff

+++



Opticon Meeting

Pyramid for co-phasing

3 DoF per segment 64x64 or 16x16 signal

sampling points Interaction matrix

dimension 156x6400 or 156x400

Signals for segmentDifferential piston

Center act.

Side actuator

S. Esposito



Opticon Meeting

Closed loop simulations

rms < /20

Each realization convergesto /20 rms after 7 iterationsmaximum.

An input wavefront

Piston values ofeach segment

S. Esposito



Opticon Meeting

Modal stiffness

Actuator stiffness 0.2N/m

Local ctrl dominates

FF dominates



Opticon Meeting

Backplate FEA: dynamics

84Hz168Hz

201Hz 282Hz



Opticon Meeting

Oscillation of the ribs:3.1kHz (out of the bandwidth)



Opticon Meeting

50m example

2m diameter segments

About 600 segments

2mm gap between segments

10mm thick glass:77.5kg/segment46.5tons total

INAF-Osservatorio di Arcetri 3-4 May 2002. Tuorla Observatory Opticon Meeting Adaptive secondary and primary mirrors for ELTs A. Riccardi 1, G. Brusa 1,2,

Documents