Change page style: 

Altair FPRD Summary

You are here

Below is a summary of the Altair specifications, as in the Functional and Performance Requirements Document.

Altair is the natural guide star adaptive optics system for the Gemini-North telescope built at the Herzberg Institute for Astronomy (HIA). 

The science performance requirements for Altair include:

  • Wavelength Coverage
    1-2.5 microns baseline with requirement of changeable dichroic, allowing observations to 0.85 microns. A goal is to extend observations to 5 microns and not preclude use to 0.5 microns.
  • Throughput and Emissivity
    The deployment of AO will not lower the telescope throughput by more than 25% over the baseline wavelength range. With the changeable dichroic the same throughput specification is required for 0.85 - 1.0 microns. The total emissivity of the telescope and AO system (without AtmDC) in K must be less than 19%, with a goal of this emissivity out to 5 microns.
  • Flat Fielding
    It is required that the flat field instability does not cause systematic effects larger than the photon noise over a 5x5 arcsec2 region within a 1 hour integration in J through K. It is a goal to match this in L and M.
  • Strehl Ratio
    A minimum on-axis Strehl ratio delivered to the detector of the near-infrared imager (NIRI) of 0.4 in H during median seeing conditions for bright guide stars. The on-axis strehl must be met, during any one hour of exposure, within 45 degrees of zenith, when atmospheric conditions remain constant. The intent is that, when applied to a background-limited point source observed at a zenith angle of 45, this Strehl ratio will result in roughly a factor of 2 improvement in S/N ratio over what would be obtained with tip/tilt/focus compensation during the same integration time. This system is expected to deliver a commensurate performance at J.
  • Field of View
    2 arcmin diameter unvignetted field of view. [Conjugation of the DM to an optimal altitude (6.5 km) is required. ]
  • Sky Coverage
    The intent is that sky coverage be maximized for the specified strehl ratios. The natural guide star AO system should be designed such that it can be upgraded to a laser guide star system and have a flexible control architecture, with the priority to increase the system's sky coverage at the above performance levels.


Sections of the workscope requirements expand on, orclarify ambiguities in, the science performance requirements:


  • "If required, it is acceptable to manually change a dichroic to allow scientific observations below 1 micron or above 2.5 microns, said change(s) being a daytime activity.
  • It shall be possible to execute a dithering sequence with Altair and a scientific instrument. Any changes internal to Altair as a result of executing a "dithering" sequence while science observations are done in the K band shall not result in an illumination change at the science array of more than one part in 100,000 of the background illumination. It is a goal that any changes internal to Altair as a result of executing a "dithering" sequence while science observations are done in the L or M band shall not result in an illumination change at the science array of more than one part in 3,000,000 of the background illumination. It is a goal to not move or otherwise change any surface in the science optical path during the "dithering" sequence. The intent of this requirement is to limit systematic effects to be less than photon noise over a 5x5 arcsec^2 area and a 1-hour integration.
  • Altair shall have a "minimum variance reconstructor" with adjustable bandwidth that varies automatically during the science observation. There should also be provision to optimize the reconstructor matrix frequently during an observation."
  • NRC and AURA agree that the requirement for laser guide-star upgradability can be satisfied for PDR by providing a conceptual opto-mechanical design with either:
    • A single WFS which can be moved between the position to acquire natural guide stars, and the position needed to track the focus motion of laser beacons.
    • A second WFS, fed after the NGS pickoff, which can track the focus motion of a laser beacon.


The following items are highly desirable ("workscope goals"), although they are not included in therequirements:


  • Reduction of the total emissivity of Altair, without the AtmDC in the adaptive optics beam, to less than or equal to 10% in the K band. The intent is that the total emissivity of the telescope and Altair (without the AtmDC) in K should be less than 14%.
  • Implementation of more sophisticated control algorithms, such as adjustable modal control, with the intent that, for science objects which do not have sufficiently bright guide stars, the number of degrees of correction can be reduced to improve performance.
  • Development of image processing algorithms including, but not limited to, the reduction of scientific data and calculation of point spread functions.
  • Provision of a remotely changeable dichroic mechanism.



3. System Overview

3.1 Introduction

The basic goal behind the Altair design is to produce an effective, competitiveinstrument for image correction for near-infrared astronomy. The design was createdthrough an integrated approach involving mechanical, electronic and optical constraintsand optimization. Careful attention was paid to maximizing sky coverage with natural guidestars while meeting the top level performance requirements noted next. In practice thismeant preserving guide star photons, sometimes at the expense of emissivity. Yet for equalsignal to noise, Altair will reduce exposure duration by factors of 2-4 for typicalobservations. Its superb image resolution will enable important astronomy, impossible byany other means on Gemini. Observing efficiency has as much to do with ease of calibrationand use, stable mechanisms and rapid configuration and guide star acquisition as withtransmission of the optics. One novel aspect of Altair is that it will double theisoplanatic patch size by conjugating the DM to the optimal altitude. This has importantimplications for sky coverage, laser guide-star performance and implementation.

To allow for a laser guide star upgrade, dead mass, space, and spare card slots forelectronics have been included in the mass, power and center of gravity analyses. Forthese calculations, we assumed the worst case of an additional motorized laser beacon WFStogether with extra processing electronics for this WFS. We made similar allowances incase the Gemini Acquisition and Guide unit does not meet its goal of processingon-instrument wavefront sensors rapidly enough.

3.2 Capabilities of Baseline Design

Altair duplicates the original telescope beam, except of course the image quality isbetter. Altair preserves the f/16 focal ratio, reproduces the location of the focal plane,and retains the exit pupil at the telescope secondary mirror. As well, Altair flattens theoutput field, whereas the raw telescope beam has a two metre radius of curvature. Theunvignetted field of view is 2 arcmin diameter. While its fast tip/tilt mirror isconjugate to a telescope pupil, Altair' deformable mirror (DM) is at an image of the layer6.5 km from the telescope. This increases sky coverage with natural guide stars andameliorates focus anisoplanatism with laser beacons.

There is a deployable I,J,H band atmospheric dispersion compensator in the sciencepath, downstream of the beamsplitter.

A calibration source is deployable at the input focus, upstream from all Altair optics.Together with the input and output shutters, this source will be used for daytimecalibration of Altair.

One beamsplitter covers the 0.85 to 2.5 micron minimum requirement. While manualchanges of a dichroic are scientifically acceptable to meet the goals for extendedwavelength range, the baseline does have room for a two-position beamsplitter changer.Whether it is installed depends on the entire detailed costing of Altair, during the nextphase. A single WFS foreoptics path and detector can handle both types of guide stars byremotely changing a lens group and a lens.

3.3 Physical Packaging and Opto-Mechanics

Altair is a Cassegrain mounted facility on the northern Gemini 8 m telescope on MaunaKea. It is designed to accept the telescope f/16 beam, correct aberrations, and return itto the science instruments as transparently as possible -- the instruments do not have tobe reconfigured when Altair is used. Altair recreates the exit pupil, focal ratio andfocus location of the original telescope beam.

All instruments on Gemini mount on the Instrument Support Structure (ISS), a rotating1.6 m cube below the primary mirror. At the centre of this cube, a moveable Science Foldmirror with three degrees of freedom directs the telescope beam out the central hole oneach side face of the cube. All instruments must accept a beam that focusses 300+-1 mmpast the centre of each face of the ISS.

Alone among current AO projects, the Gemini Adaptive Optics System has its deformablemirror located at an image of the atmosphere far above (6.5 km) the telescope. Racine andEllerbroek have shown this will increase sky coverage and improve performance.

3.4 Control system

Altair uses a 177-element continous facesheet deformable mirror and a 12x12Shack-Hartmann wavefront sensor based on a CCD-39 from EEV (3e- noise at akilo-frame/second sample rate). In addition, Altair also has a very strong goal to acceptOIWFS T/T/F signals, for operation with laser guide stars. The AO control system processespixels from its own WFS in the "reconstructor" and blends the result with theT/T/F signals from the OIWFS. The combined measurements drive the tip/tilt and deformablemirror in Altair at sample rates up to 1.5 kHz.

To preserve the maximum dynamic range and small signal bandwidth on these mirrors,Altair offloads their time averaged tip/tilt and focus positions to the secondary mirrorvia the synchro bus at 200 Hz. Similarly, the first ten low-order quasi-static aberrationsare sent via an ethernet to the telescope control system which corrects the telescopedrive rate, the primary mirror figure, and the XY collimation of the telescope.


4. LGS upgrade path

From an opto-mechanical viewpoint, the laser guide star upgrade path involves changesto the optics in the wavefront sensor optical path. The requirement is to be able toswitch between LGS and NGS modes remotely. The optical design for the LGS path modifiesthe field lens and the collimator optics. The field lenses for each design can be mountedon linear stages and moved independently in and out of the beam. The collimator lenses aremore difficult as they occupy the same region in the optical design and they each requireadjustment as the system operates. The NGS collimator moves in focus and the LGScollimator moves in focus and zoom to accommodate different laser altitudes. A mechanismis required to provide all these motions and to interchange the LGS and NGS collimators.

The current optical layout of the wavefront sensor path may have to be modified inorder to leave room for a mechanism to change between the LGS and NGS collimators.Currently, the NGS WFS collimator lens barrel is close to the science path collimatormirror leaving little room for a changing mechanism. Moving the WFS collimator lens barrelback further from the science collimator mirror would provide room for the mechanism. Thiswould be accomplished by modifying the fold and gimbal mirror angles to move the WFScollimator lenses back further.

When using laser guide stars (LGS) the WFS foreoptics must focus the spot of light fromthe sodium layer onto the AOWFS CCD, while keeping an image of the DM consistently mappedonto the lenslets. The problem is that the range distance to the sodium beacon constantlychanges with weather and telescope zenith angle. To account for this, the laser beaconcollimator has two extra independent axial motions for lens elements. This keeps the DMimage focussed at constant size on the lenslets, while focussing the laser spots onto theCCD.

Another important upgrade provision results from a feature of the Gemini telescopes --all instruments operating shortward of 5 microns have on-instrument wavefront sensors(OIWFS). These 2x2 Shack-Hartmann OIWFS can measure tip, tilt and focus and therefore mayhave in three roles with natural guide stars:


  • fast guiding the telescope
  • slow flexure reference for Altair
  • It is a strong goal to use these OIWFS as a high-speed natural guide star tracker (as well as flexure reference) for Altair.


If the OIWFS fills all three roles, it is already capable of serving as the naturalguide star tracker required in conjunction with a laser beacon. A laser beacon makes atwo-way trip through the atmosphere, but celestial light makes a one way trip. Thereforethe LGS is insensitive to tip-tilt caused by atmospheric turbulence. If the outbound lightbends sideways, the return light will follow the same path down and arrive at thetelescope on-axis. The laser beacon will not appear to wander. However, inbound starlightwill be bent off-axis, and tip-tilt errors will result. The natural guide star tracker(i.e. OIWFS) measures atmospheric tip-tilt and sends the data, via the Acquistion andGuide (A&G) system, to Altair.

Similarly, because the range distance to the laser beacon varies unpredictably, theAOWFS guiding on the LGS cannot detect long term absolute focus. However, the AOWFS canmeasure instantaneous deviations from the current average focus, and drive the DM based onthese high speed focus signals, providing it relies on the OIWFS for an absolute focusreference for the low temporal frequencies of focus.


5. Performance Simulations

5.1. First-order Calculations

Brent Ellerbroek has calculated the performance of the Gemini baseline AOS under avariety of conditions.



  • The simulations do not take into account uncorrected aberrations from the telescope or focal-plane instrumentation.
  • The simulations consider tip/tilt and higher-order aberrations seperately. References to HO means the effects of the tip/tilt are not included. Residual jitter can arises from uncorrected atmospheric turbulence as well as telescope wind shake. For these calculations, the telescope wind shake power spectrum has an assumed form.


The baseline assumptions in these simulations are tabulated below.

Parameter NGS HO LGS HO Tip-Tilt
WFS Subapertures 12 x 12 12 x 12 2 x 2
Sensing Band (microns) 0.4-0.8 0.589 J + H
Atmospheric Trans. 85% 85% 95%
Total Efficiency 34.4% 34.4% 34.4%
Read Noise 3 3 6
Sky Background 332 0 6890
WFS Pixel Scale (arcsec) 0.250 0.250 0.170
Sampling Rate (Hz) 10-1000 10-1000 10-1000

N.B. The abbreviation HO stands for high-order and generally refers to calculationsexcluding tip/tilt errors.

Strehl Ratio as a function of Guide Star Distance

A comparison of the Strehl ratio as a function of the distance to the natural guidestar. In the case of the NGS AOS, the natural guide star is used for all orders ofcorrection so this plot can also be taken as the variation of the Strehl ratio within thefield of view. Note that the difference between the altitude conjugate and pupil conjugateis nearly a factor of two in radius. In the case of the LGS AOS, the natural guide star isonly used for tip/tilt correction. This is the principle reason the LGS AOS depends lesscritically on the distance to the guide star. Unlike the NGS case, this plot should not betaken as an indication of the variation of the Strehl ratio across the field of view for aLGS AOS. The plotted Strehl is calculated using only the tip/tilt angular isoplanatismterm. Again, in both cases, the Strehls do not take into account residual image motion,nor residual static aberrations from the telescope or focal-plane instrumentation.

Sky Coverage NGS & LGS

A comparison sky coverage for NGS and LGS adaptive optics systems. The upper threecurves are the LGS AOS sky coverage. In the calculation median seeing conditions areassumed, the zenith angle is zero, and the deformable mirror is conjugate to 6.5 km.

Slit Throughput as a function of Residual Jitter

Here Brent has plotted the variation of the power through a spectrograph slit as afunction of the residual image motion (tip/tilt). The spectrograph slit width is set totwice the diffraction-limit at that wavelength. The 'Total Strehl Ratio' now includes theeffects of residual tip/tilt errors. The right-most point on each of the curves representsthe throughput if the residual image motion is zero. Increasing residual image motion isto the left on the plot. Note that the curves plateau because while jitter lowers thetotal Strehl ratio, it does not effect the slit throughput until the residual jitter isroughly equal to the slit width. Again the calculations do not consider the uncorrectedaberrations in the telescope or spectrograph up to the slit.

Page compiled by Mark Chun. Last update Aug 23, 2012; Chad Trujillo

Gemini Observatory Participants