21 November 2017 Development of a three-mirror anastigmat telescope for the GERB experiment
Author Affiliations +
Proceedings Volume 10569, International Conference on Space Optics — ICSO 2000; 105691W (2017) https://doi.org/10.1117/12.2307913
Event: International Conference on Space Optics 2000, 2000, Toulouse Labège, France
The GERB experiment, on-board Meteosat Second Generation, aims at monitoring the Earth radiation budget within a broad spectral range (0.32 –30 ‡m). This paper outlines the development of the GERB imaging subsystem, a f/2 three-mirror anastigmat telescope with a 18° x 0,28° rectangular field-of-view. The telescope is an all-aluminium design, comprising a primary off-axis elliptical mirror and two spherical ones, with a largest size of 100 mm. After integration and environmental testing, its global on-axis imaging performance reached 0,45 ‡ rms at 633 nm for an optical design value of 0,27 ‡ rms. The global opto-mechanical tolerance analysis of the design phase defined an integration sequence able to keep the individual alignment of each mirror within the accuracy needed to ascertain the whole telescope quality.



Understanding the behaviour of the Earth climate is one of the most important scientific problems to date that man has to face.

The point is to gain a deeper insight into the processes that control the climate system variability and global balance as well as to assess the consequences of man’s activities.

Satellite-borne equipment is particularly well suited to collect relevant global data in numerous complementary fields and to exploit the associated synergy.

In particular, the GERB (Geostationary Earth Radiation Budget) experiment is to fly on MSG (Meteosat Second Generation) satellite and aims at measuring the balance between the incoming radiation from the sun and the outgoing reflected and scattered solar radiation plus the thermal infrared emission to space.

The geostationary orbit allows to get a very good temporal sampling of important diurnal processes affecting clouds and water vapour and brings a perfect complement to polar orbiting measurements.

The GERB radiometer will provide data for the region of the globe covered by MSG:

Short-wave (0.32-4 μm) and total (0.32-30 μm) measurements will be performed, the long-wave (4-30 μm) data being obtained by subtraction.

The GERB instrument is divided into three subsystems:

  • the Instrument Optics Unit (IOU) comprising the optics, the detector and the calibration sources.

  • the Instrument Electronics Unit (IEU) for data handling, power supply and thermal control.

  • the Mechanisms Control Electronics (MCE) for quartz filter and despin mirror

The instrument characteristics are:

  Spin-stabilised (100 rpm)
Spectral bandsShort-wave0.32 μm - 4 μm
 long -wave4 μm - 30 μm (by subtraction)
 total0.32 μm - 30 μm
Short-wave filterquartz 
Field of view (pixel size)44.6 × 39.3 kmat nadir (NS × EW)
Telescope3-mirror anastigmat 
 + 1 fold mirrorto minimise polarisation effects
 + 1 descan mirrorto remove the satellite rotation
Detector256 × 1 thermallinear array in North-South direction
Calibration sourcesthermal infraredblackbody
 solarintegrating sphere
Sampling time300 secondsfull Earth disc, both channels
Mass25 kg 
Thermal Control18°C - 22°Cat telescope level

The GERB instrument is produced by a European Consortium led by the UK, with the involvement of Belgium and Italy.


  • Rutherford Appleton Laboratory (RAL) main contractor

  • Imperial College (ICSTM) science and calibration

  • Leicester University detector and signal conditioning electronics

  • Hadley Centre science


  • RMIB fluxes, near real-time data products

  • AMOS/OIP telescope assembly


  • Officine Galileo descan mechanism / quartz filter mechanism



The imaging subsystem of GERB is a three-mirror anastigmat (TMA) telescope. The all-reflective design is well adapted to the broad useful instrument waveband. The requirements were:

  • optical configuration (see fig. 2.1.)

    entrance aperture stop :- size 20.6 mm × 20.6 mm (f-number : f/2)
    mirror M1 (PM)- off-axis concave ellipsoidal (parent shape: prolate ellipsoid)
     - rectangular sizes: 57 mm × 30 mm
    mirror M2 (SM)- convex spherical
     - rectangular sizes: 30 mm × 11 mm
    mirror M3 (TM)- concave spherical
     - rectangular sizes: 106 mm × 74 mm
    field-of-view :- 18° × 0.28° (North-South bands on Earth disc)

  • global imaging performance :

    RMS telescope wavefront error in waves (λ = 633 nm)

    in-field(edge of the FOV)0.44λ0.80λ1.00λ

    The design values are obtained after design optimisation. As-built values include contributions of manufacturing and alignment.

    In-flight values represent the performance awaited from the telescope after launch, encompassing effects such as launch vibrations, thermal cycling and centripetal loads due to satellite spinning.

  • straylight performance

    The straylight is controlled by telescope baffling performed by RAL and by minimising the mirror surface roughness

    • for spherical mirrors : Rq < 2.7 nm RMS

    • for aspherical mirrors : Rq < 5 nm RMS

    Mirror cosmetic quality is imposed to be 60-10.

    Contamination control :

    • particulate contamination level < 550 ppm

    • molecular contamination level < 10−7 g/cm2

  • mirror reflectivity : UV- enhanced protected silver coating

  • telescope focal length : 1% tolerance with respect to nominal design

  • boresight error : <70μm

  • telescope interface with optical bench : 0,2 % tolerance on mirror locations with respect to nominal design

  • mass budget : 2 kg

  • environmental requirements :

    • thermal requirement :

      Telescope modeOperatingNon operating
      Telescope temperature18°C22°C−30°C50°C

    • launch environment : sine and random (15 g RMS) vibrations

    • centripetal loading

The GERB instrument being located at 1,5 m from the satellite rotation axis (100 rpm) it will undergo a permanent centripetal acceleration of nearly 17 g.

Fig. 2.1.

GERB TMA Optical Design




The opto-mechanical tolerance analysis was performed with the purpose of establishing the specifications on individual components (mirrors and mounts) which enable the telescope to reach the system performance.

The analysis was further split into ground as-built and in-flight (launch and in-orbit environment) performance by separating the contributors to each working case.

The system performance was essentially represented by the telescope wavefront error (on-axis and in-field values) which had the major impact on GERB radiometric accuracy. A parallel analysis was also made for telescope focal length and boresight error.

By means of a global ray tracing model, inverse sensitivity tables were established between the telescope wavefront error and the identified components parameters. The following table summarise the set of accounted contributors for as-built (table 3.1) and in-flight (table 3.2) cases.

Table 3.1

Contributors identification table for As-Built WFE tolerancing

curvature radiusXXX
conic constantX--
surface accuracyXXX
Z local alignment (axial)XXX
Y local decenterXXX
X local decenterXXX
α tilt (around local X axis)X--
βtilt (around local Y axis)X--
Total :18 contributors

Table 3.2

Contributors identification table for In-Flight WFE tolerancing

surface deformationXXX
Z local alignment (axial)XXX
Y local decenterXXX
X local decenterXXX
α tilt (around local X axis)XXX
β tilt (around local Y axis)XXX
γ tilt (around local Z axis)XXX
Total :21 contributors

The next step was to build a preliminary tolerance budget based on the inverse sensibility table and on the actual capabilities in terms of manufacturing, integration and alignment of the telescope.

In the same way, constraints were put on the mirror design in order to take into account the feedback from the tolerance analysis.

Due to the inherent asymmetry of three-mirror telescopes, a statistical approach for combining the tolerance parameters is required.

The parameters can then be considered as random variables with a defined density of probability. In this way, it is possible to determine the probability to stay below a given wavefront error after telescope integration and alignment.

We took another approach, based on a worst-case analysis for both on-axis and in-field configurations. This worst case was tracked through an iterative procedure that attributed different values to the set of parameters within the defined tolerance range and looked for the maximum telescope wavefront error.

We also considered compensation mechanisms, based on one part on telescope refocusing at the detector level and, on the other part, on adjustment of mirror interdistance.

The final tolerance budget was tuned from the results of this analysis in order to stay within the required system performance.

The following table (table 3.3) provides an overview of the typical tolerances for the as-built performance:

Table 3.3

TMA Tolerancing overview

radius of curvature0,1 %0,1 %0,05 %
conic constant0,3 %--
decenter10 μm10 μm10 μm
tilt10 arcsec--
RMS surface accuracy (at 633 nm)λ/8λ/10λ/30



The concept of mirror units was selected on the basis of the manufacturing and integration plans. An all-aluminium design was retained to minimize the thermal gradients across the telescope. The mirrors were designed with integral mounts; the mounting strain path was isolated from the mirror surface by a slot located between the latter and the mounting screws, which allowed the created pad to play the role of a flexural spring.

The mirrors units were calculated to resist to the launch loads (mirror-mount assembly) and to the centripetal loading in operation (minimization of the surface tilt and deformation). Mirrors were manufactured using the diamond turning technology that also allowed to machine flat and coplanar mounting interfaces to the same tolerances as the optical surface figure.

The mirror manufacturing sequence was the following:

  • rough machining of aluminium alloy (6061-T6) blank

  • stabilisation heat treatment (temperature ageing)

  • diamond turning of optical and reference surfaces

  • optical testing

  • stabilisation heat treatment (temperature ageing)

  • optical testing

  • electroless nickel plating

  • stabilisation heat treatment

  • optical figuring to required accuracy and required roughness

  • optical testing

  • protected silver coating application



The first integration step consisted in assembling mirror and mounts (for PM and TM) and in checking afterwards the optical quality. A tightening torque in adequacy with the launch vibration levels was applied to the mounting screws. This operation affected somewhat the surface figure of TM which remained nevertheless compatible with the allocated budget.

The second step consisted in integrating the mirror units on their support baseplate. Due to the high accuracy needed, all the procedure was performed on the marble of a 3D coordinate measuring machine (CMM), installed in the cleanroom. The CMM constituted the reference frame for the telescope alignment and allowed to catch the locations of PM foci and SM and TM centers of curvature. The procedure began with the integration of the elliptical primary mirror on the baseplate: the beam emitted by the interferometer was focused on the first focus of the ellipsoid and reflected back on a ball placed at the second focus, after having hit the mirror (see fig. 5.1.)

Fig. 5.1.

Telescope during integration


This step was the most critical of the alignment and required five independent adjustment possibilities on the mirror (three translations and two rotations), realised by lapping of washers at the interface between the mount and the baseplate.

Once the primary mirror integration was completed, the residual position errors with respect to baseplate were recorded, so that the nominal positions of SM and TM were corrected accordingly. PM orientation therefore fixed the whole telescope orientation (at less than 1 arcmin of the nominal one).

SM and TM were then integrated on the basis of the corrected coordinates using the interferometer focused on the centres of curvature and in autocollimation on the mirrors. Lapping of washers at the interface between mount and baseplate allowed to complete integration. (see fig. 5.2).

Since the whole procedure was performed within the accuracy requirements for individual components, the global specification for the telescope was reached without having to make use of the computer model to find the optimized compensating adjustments.

The results of alignment were:

TMA Tolerancing overview

 As-built specification (wavefront) λ = 633 nmTest (wavefront) λ = 633 nm
on-axis0,57 λ RMS0,45 λ RMS
in-field0,80 λ RMS0,69 λ RMS

Fig. 5.2.

Telescope at the end of integration




The telescope underwent thermal cycling (from −30°C to +50°C) and vibration testing (see fig. 6.1).

Fig. 6.1.

Telescope on the shaker


There was no impact on the telescope wavefront error, while a best focus shift of 30μm was measured.

Introducing the effect of centripetal loading on the telescope quality, the final status was :

 In-flight specification wavefront λ = 633 nmTest + centripetal loading
on-axis0,77 λ RMS0,60 λ RMS
in-field1,0 λ RMS0,80 λ RMS



The design and development of an all-aluminium three-mirror anastigmat for the GERB experiment has been presented.

From the opto-mechanical tolerance analysis, a dedicated alignment procedure was defined, involving interferometry combined with accurate metrology performed on a 3D coordinate measuring machine.


We wish to thank E. Sawyer and M. Caldwell from RAL project office for the excellent spirit they brought in the project, as well as G. Brandt and L. De Vos from OIP for their collaboration. We also wish to acknowledge the contribution of the remaining people of AMOS team : S. Denis (optical computations), C. Delrez (finite-element analysis), O. Dellis (CAD), and G. Chupin (Optical workshop).

© (2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Pierre Gloesener, Pierre Gloesener, Eric Quertemont, Eric Quertemont, Carlo Flebus, Carlo Flebus, } "Development of a three-mirror anastigmat telescope for the GERB experiment", Proc. SPIE 10569, International Conference on Space Optics — ICSO 2000, 105691W (21 November 2017); doi: 10.1117/12.2307913; https://doi.org/10.1117/12.2307913


Back to Top