Scientific Objectives

Extreme adaptive optics (XAO) systems dedicated to the search for extrasolar planets are currently developed for most 8-10 meter telescopes. Extensive computer simulations have shown the ability of both Shack-Hartmann and pyramid wave front sensors to deliver high Strehl ratio correction expected from extreme adaptive optics but no experiments have been realized so far. In the frame of OPTICON JRA1, the high order testbench (HOT) implements extreme adaptive optics on a test bench with realistic telescope conditions reproduced by star and turbulence generators. A 32×32 actuator micro deformable mirror (DM),one pyramid wave front sensor, one Shack-Hartmann wave front sensor (SHS), the ESO SPARTA real time computer and an essentially read-noise free L3 CCD60 provide an ideal cocoon to study the different behavior of the two types of wave front sensors in terms of linearity, sensitivity to calibration errors, noise propagation, specific issues to pyramid or Shack-Hartmann wave front sensors, etc. Among the large variety of tests, the following topics will be studied:

  • System calibration, interaction matrix: Effects of mis-calibration
  • Noise propagation
  • Linearity, linear range and cross-talk
  • Misregistration. Effects of mismatch between DM and WFS
  • Aliasing error. Effects of high spatial frequency turbulence on WFS measurement.
  • PSF characteristics & residual aberrations
  • Chromatic effects
  • Extended object for wavefront sensing.
  • erformance (Strehl, sensitivity, robustness)
  • Special PWS issues: modulation, diffractive regime, impact of pupil shape
  • Special SHS issues: spatial filtering
  • >Coronagraph concepts
  • Woofer/Tweeter concept (2 DMs – 1 WFS, relevant for deformable secondary mirror control with active optics and for an ELT)

All these tests will allow us to better understand all the limitations which can reduce performance of a eXtreme AO instrument on 8-m class telescopes like SPHERE and even the corresponding instrument at an ELT like EPICS.

Instrument Description

Instrument Design

Figure 1. HOT setup on the MACAO test bench

Click to enlarge the Top View

MACAO test-bench

The MACAO test-bench was initially developed for the assembly, integration and testing of the ESO multi application curvature AO (MACAO) systems for the VLTI. Reference sources can be chosen from a number of optical fibers with variable diameters mounted on a wheel. Dynamical optical aberrations are generated by a source module with two rotating phase screens which produce an f/16.8 beam. Seeing values of 0.5" and 0.85" with attenuated low orders resembling first stage correction by a 60 element mirror, and 0.65" with full Kolmogorov turbulence can be produced. A spherical mirror on axis transforms this beam in a converging f/51.8 beam. A flat mirror with a central hole simulates the VLT central obstruction (14%). The pupil plane, where either a flat mirror or the 60-element MACAO bimorph DM can be installed in a tip tilt mount (TTM), is located at about 1010 mm above the table level. Pupil Masks to simulate different pupil shapes can be installed close to its surface. This design provides a field of view of 10" in diameter and a diffraction limited image quality with a wavefront error 46 nm rms WFE has been measured.

HOT common path

Following the MACAO test bench optics, the beam is folded on the table and feeds the HOT optics with the 32x32 actuator MEMS DM (MDM, Boston Micromachines). A 10.5 mm diameter pupil is imaged on the micro-DM by a spherical mirror (f = 500 mm). Another spherical mirror, identical to the first one, produces an f/50 telecentric beam. After the second spherical mirror, the beam is intercepted by a dichroic transmitting the infrared light to coronagraph and infrared camera while the visible light is directed towards SHS and PWS. The optical quality by design is 180 nm PTV (~40 nm rms), so that the DM dynamic range can almost entirely be used to correct for turbulence.


The Shack-Hartmann wavefront sensor (SHS), developed by the university of Durham, is based on a 31x31 subaperture lenslet array and an ANDOR iXon L3CCD60 with 128x128 pixels. A anti-aliasing filter can be inserted in the front focal plane. The SHS can be equipped with two different lenslet arrays providing pixel scales of 0.25" or 0.5" for the 4x4 pixel subapertures. The SHS real-time computer (RTC) is an all-CPU architecture.

It will in large parts comply with and work as a testbed for the ESO AO real-time computer platform SPARTA. The RTC controls the MDM and the TTM via control electronics produced by Shaktiware whose interface will be initially based on Ethernet. For faster operations at the frame-rate limit set by the iXon camera of around 450Hz, SPARTA-HOT RTC can easily be upgraded with a serial FPDP card, once the ShaktiDrive is also upgraded to the same interface. At a later stage the RTC will also be able to interface to the MACAO DM in order to test the woofer/tweeter concept.


The Pyamid wavefront sensor (PWS), build by Arcetri, will have two possible pupil sampling obtained changing the final camera lens namely a low sampling mode with 31x31 subapertures and a high sampling mode with 48x48 subapertures. Any other intermediate sampling between these two could be achieved using a proper camera lens. As the SHS, the PWS is based on the ANDOR iXon L3CCD60. Wavefront modulation needed for system characterization and sensing optimization is obtained using a fast steering mirror conjugate to the system pupil. A refractive double pyramid design is used that ensures efficient alignment and sufficient suppression of chromatic effects. The PWS uses a dedicated all-CPU RTC which controls the MDM and the TTM via the ShaktiDrive with the Ethernet interface.

Coronagraph and science camera

At least three different types of coronagraph (classical Lyot, pupil apodized Lyot and phasemask) will be implemented in HOT und studied by experiment in collaboration with the Paris Observatory in Meudon. The main focus of this study will be on coronagraph performance in the presence of positioning errors (image and pupil, may be chromatic), alignment (pupil telescope vs pupil instrument, translation + rotation), chromaticity, and AO residual image quality. Images will be recorded in different NIR bands with the RASOIR infrared camera using up to one quadrant (512x512 pixels) of a HAWAII 1k array and a pixel scale of ~10mas/pixel.

Baseline Specifications

Frame rate
Pupil masks
Light source
Field selection

1024 act. MDM,
60 act. BIM,

31x31 SHS and PWS,
L3CCD60 with 128x128 px

2 x all-CPU
~100Hz (TBD)
white light, variable size
0.5", 0.65", 0.8"
~0.5" PTV (TBC)

Documentation and Experiment results

All the major results can be find in the following reports and articles: