Dealing with turbulence in MCAO Roberto Ragazzoni INAF – Padova (Italy) Alghero, Sept. 17 th 2008 Astronomy meets Metereology Dealing with turbulence in.

Dealing with turbulence in MCAO Roberto RagazzoniINAF – Padova (Italy)

Alghero, Sept. 17th 2008Astronomy meets Metereology

Dealing with turbulence in MCAO

Roberto Ragazzoni, Yazan Almomany, Carmelo Arcidiacono, Renato Falomo, Jacopo Farinato, Marco Gullieuszik, Osservatorio Astronomico di Padova (Italy);

Emiliano Diolaiti, Matteo Lombini, Osservatorio Astronomico di Bologna (Italy);Alessia Moretti, Osservatorio Astronomico di Padova (Italy);

Giampaolo Piotto, Univ. degli Studi di Padova (Italy);Enrico Marchetti, Robert Donaldson, European Southern Observatory (Germany);

Roberto Turolla, Univ. Degli Studi di Padova (Italy)



What is MCAO…??? Why is important the knowledge of the 3D turbulence???



Telescope

DM1 DM2

Turbulent layers#2 #1

Atmosfera

WFS

UP

Multi Conjugated Adaptive Optics: MCAO



An historical perspective…

• MCAO is a “vision” of J. Beckers in 1988

• Pyramid WFS envisaged in 1995

• Layer Oriented approach is shown in Backaskog 1999 and Munich 2000

• LO approved as part of MAD in 2001

• MCAO with 3SH and LO

are on the sky in 2007!



Canary, fall 2001



Why pyramid..?

• They take advantage of the full aperture (spot is diffraction limited on the pin of the pyramid) Better limiting magnitude

• Better aliasing (larger Strehl in HO mode)

• It is a pupil plane WFSensor!

• It is “cheap” in terms of pixels usage





Prototyping(4 stars)

Opticaldesign

Optomechanicaldesign

Glass andAlluminum



0,00

10,00

20,00

30,00

40,00

50,00

60,00

70,00

7 8 9 10 11 12 13 14 15 16 17

V Magnitude

Str

ehl R

atio

(%

)

400 Hz - 0.85"

200 Hz - 0.85"

100 Hz - 0.85"

50 Hz - 0.85"

400 Hz - 0.45"

200 Hz - 0.45"

100 Hz - 0.45"

50 Hz - 0.45"

Laboratory test in closed loop give evidence of higher limiting magnitude for the Layer Oriented WFS w.r.t the Star Oriented ~1.5 magnitude fainter

LOWFS SHWFS

vWFS on MAD vs 3SH…: FoV of single stars enlarger ~0.95arcsec!





The MAD-LO run• Basically the whole group of LO-MAD plus

ESO support (3 initially, then 1)

• Total of 9 contiguous nights

• First 3 night of technical run under ESO responsibility

• Then 6 nights of “GTO” basically devoted to science







70.5”



• 107x107arcsec

• GLAO

• 5 NGSs

• Vmag tot=13.8

• Here is S=10%

• Var. +/- 3%











Not a star!!!



A. Moretti et al.





• We think to have material for further 3 science papers on other objects…

• A very productive run indeed, but….

• Which is the achieved quality compared to the SO one?

• Which is the sky coverage one can compute given the actual performances?

• Any answer will be plagues by doubts about seeing, f_G, scaling laws…



A. Moretti et al.

Omega Cen (public data)NGC6388 (LO data)





LayerOriented

Star Oriented

Layer oriented

Star Oriented

With an uniform “standard” reduction magnitude errors are slightly (10%) betterfor LO than SO, and viceversa for centroiding





0.45”

0.60”

0.95”



3x V=11.5



3x V=12

3x V=13



3x V=12

3x V=13

~1.5mag







Red: estimates given in the 2002 paperBlue: left – actual measurements in the sky; right – laboratory measures



Red: estimates given in the 2002 paperBlue: left – actual measurements in the sky; right – laboratory measures



Remember that the CCD is blue-sensitive…!!!

0.70 … 0.99 0.10 … 0.25 0.07 … 0.25

!



Areas to make MCAO with NGSs more sensitive…

• Using the Taylor hypothesys

• Layer Oriented “sees” the layers of turbulence so can use “several” previous measurements to estimate the coming one (not prediction, but better SNR of WFS…

“Frozen” layer2000



Areas to make MCAO with NGSs more sensitive…

• Using non linear co-addition of stars

• It is well known that enlarging the FoV do not improve the correction because it is lowered the thickness of turbulence sampled… if linearly coadded!!!

a

2001



Linear or non-linear…???

Layer you are “looking” (say 8km)

a few stars





a lot of stars





a few stars

This layer is sampled Ok…





a lot of stars

This layer is smoothed out!!





a lot of stars

+ + +

Numerically averaging leads to the same trouble!





a lot of stars

multiply or adding the square

It only goes to zero if all the WF is flatted…





a lot of stars







a lot of stars







a lot of stars





Propagation noise is more complex…

1st DM

2nd DM

Pupil plane



Non-linear behaviour…• Algorithm is not conventional as the “sign” of

correction is unknown• A toy-model using the “bang-bang” algorithm

gives a bandwidth that is at worst 2x less…• There is a geometrical limit, that is about 10’ for

a 42m telescope (and 25’ for a 100m one…)• This makes the gain in FoV of 25• So the overall gain is in the ballpark of 12..!!!!• One order of magnitude improvement in sky

coverage!!!



Future of Layer Oriented• A LO instrument started in 2002 now would

produce big science• NIRVANA is not optimized for deep single arm

MCAO but, ok…• In the timescale of a new instrument LGSs

should prove their maturity (or not???)

• Unexplored ways to improve sensitivity include Taylor hypothesis usage and non-linear coaddition of reference stars

• Mixing up LGS and NGS is a safe, robust way, and there are unexplored new schemes for that



Metapupil

4 NGSs…and 2 LGSs!

Pupil planePupil plane Focal plane

NGS

LGS



Conclusions…

• Beating 3D turbulence, after 20yrs from the J”D”B is a reality…

• MCAO is in its infancy…

• There are unexplored realms where NGSs based MCAO can achieve larger sky coverage…

• LGS are being deployed on a few sites and will progress as well…

• Stay tuned for a turbulent future!!!

Dealing with turbulence in MCAO Roberto Ragazzoni INAF – Padova (Italy) Alghero, Sept. 17 th 2008 Astronomy meets Metereology Dealing with turbulence in.

Documents

mcao slide

metereology slide

padova italy enrico

padova italy emiliano

padova italy giampaolo

d turbulence

science slide

alluminum slide