Digital Control of a Michelson Interferometer Testbed Demonstrator for a Satellite Telescope

Author(s):
F. Musso - ALCATEL ALENIA SPAZIO
F. Bresciani - ALCATEL ALENIA SPAZIO
L. Bonino - ALCATEL ALENIA SPAZIO
S. Cesare - ALCATEL ALENIA SPAZIO

Industry:
Aerospace/Avionics

Products:
Data Acquisition, LabVIEW, Real-Time Module

The Challenge:
Realize a digital control system for a laboratory demonstrator of the co-phasing system of EUCLID space telescope. The purpose of the control system is to maintain the optical path differences among the telescope arms under 10 nm at 1 σ, a mandatory condition to ensure nominal satellite operations. This activity has been performed in the frame of the EUCLID CEPA RTP 9.9 contract commissioned by the Western European Armaments Organization (WEAO) Research Cell.

The Solution:
The control algorithm has been written in C++ language and embedded in a dynamic-link library. Control algorithm has been interfaced with NI LabVIEW using a Call Library Function Node to exchange data (measures from ADC and commands to DAC) with NI DAQ board.

Summary
EUCLID space telescope is an interferometric instrument optimized for the high-resolution optical surveillance, from a geostationary orbit, by means of the synthetic aperture technique. In order to obtain the desired co-phasing, and then the desired resolution, a complex metrology and control system is needed to ensure the necessary stability of the optical configuration. A demonstrator (called MIT, Michelson Interferometer Testbed) has been integrated in order to validate two very critical systems of EUCLID Space Telescope for achieving the cophasing condition and maintaining the fringe
pattern stabilized in the Michelson interferometer instrument. On the first part of the paper an overview of Euclid Space Telescope will be presented, then a short description of MIT performances and goals achievements.

Article
The EUCLID space telescope

Multiple Apertures Telescope configuration like the EUCLID Space Telescope one provide a unique opportunity to achieve very large aperture optical systems. The motivation for developing multiple independent telescope apertures is to provide high resolution observation from space avoiding the practical limitations in the areas of large optics fabrication (and weight) and adaptive wavefront control. Multiple telescope optics can be much smaller than large monolithic mirror of equivalent diameter having “as first consequence” an improvement on the weight and encumbrance to be launched. Michelson Interferometer with Fizeau-type combination optical configuration has been selected to implement the synthetic aperture technique. The telescope configuration consists in an array of eight sub-telescopes and a beam combining telescope at the centre of the array which collects the light incoming from the subtelescopes and produces the interferometric image on a focal plane. An optical delay line equalizes the paths of the incoming wavefront from each sub-telescope to the focal plane where they are superimposed. An interference
fringe pattern is formed on the focal plane with a good visibility when the Optical Path Difference (OPD) between the interferometer arms is kept within a fraction of the coherence length. As far as the OPD increases, the fringe pattern becomes more and more degraded, i.e. its visibility decreases. This is related to the fact that the interferometer does not operate at a single wavelength, but on a finite spectral band.In order to form a fringe pattern with a good visibility, the Optical Path Length (OPL) of the light beams travelling in the eight arms of the Michelson interferometer must be equalised
within a fraction of the coherence length of the operational spectral band. For a proper operation of the Michelson Interferometer, the OPLs of the light beams travelling in the eight arms must be equalized within 100 nm. When this condition is achieved, the interferometer is “co-phased”.After the co-phasing condition is achieved, the telescope is ready to perform the observations. During the image integration time of the focal plane, the OPD between the eight arms of the interferometer must be controlled within a fraction of the observation wavelength (i.e. OPDij < 10 nm), in order to avoid fringe “jumps” or significant variations of fringe pattern phase with consequent losses of contrast in the resulting image. If such a case occurs during an observation, the resulting interferometric image
would be completely blurred and the information necessary to reconstruct the original image of the target would be lost. This interferometer is equipped with laser metrology systems, for the measurement of the optical path difference (in absolute and relative terms) between the interferometer arms, and with a motorized delay line for the control of this optical path difference. A control system
elaborates the measurement of the laser interferometer and sends command to the delay line.Laser interferometry is by far the best technique to measure long distance variations. Several interferometric schemes are possible, but all of them are based on the interference principle: two or more light beams are generated by the same source, run different length paths and eventually are recombined (summed) on a detector which measures the intensity. The intensity on the detector is a function of the relative phase of the interfering beams which, being waves, can interfere either constructively or destructively. From the analysis of the interference signal one can get information about the path difference between the light beams.
For measuring the variations of the length between the two arms of an optical interferometer, the choice naturally falls on a laser interferometer of Michelson type. The interferometer includes two types of laser metrologies:
• An absolute metrology system (developed by INETI institute, Lisbon, Portugal), providing the actual value of the optical path unbalance between the interferometer arms, with lower resolution;
• a relative metrology system (developed by Alcatel Alenia Space Italia, Turin, Italy), providing the variation (from a given initial value) of the optical path difference between the interferometer arms, with higher resolution.

Both metrology systems are optically interfaced with the optical interferometer prototype, and are electrically interfaced with the control system that elaborates the commands for the delay line.
The role of the absolute metrology is to support the achievement of the Michelson interferometer co-phasing, consisting in the equalisation of the optical path lengths of the various arms within a fraction of the coherence length, so that a good visibility fringe pattern is formed on the focal plane of the instrument. The relative metrology provides the measurement of the OPL variations, starting
from a given initial value (the one achieved at the end of the co-phasing operation), that will be used by the control system for freezing the fringe pattern (OPD ≤ 10 nm) through the fine stage of the motorized delay line. Relative metrology is based on a Michelson interferometer metrology and has a nanometer level resolution. The OPD disturbances to be compensated during the target observation, derive from the deformations of the satellite structure under variable thermal loads and vibrations generated inside the satellite (for instance by the attitude control system) and propagating till the
interferometer mirrors through the instrument structure.

The co-phasing system laboratory demonstrator
Co-phasing system is the most critical aspect of the telescope design. In order to test and demonstrate the co-phasing system concept, i.e. 14 equalize the OPL between the interferometer arms by acting on a one degree of freedom delay line, a laboratory demonstrator has been implemented. The MIT demonstrator, consists of a simplified, laboratory-sized optical interferometer prototype, realized with the same optical configuration typology of the highresolution satellite telescope. Since the concept of the co-phasing system is to control the OPD variations between the telescope arms, the laboratory demonstrator of the co-phasing system is equipped of one control delay line (CDL) that act on one (th
primary) arm, tracking the OPL variation of the other arm (the secondary) and one disturbance delay line (DDL) that act on the secondary arm introducing an OPL disturbance with the same Power Spectral Density (PSD) as the foreseen disturbance PSD present on the satellite telescope. The performance of the laboratory demonstrator shall be the same required for the satellite telescope.Control delay line is made by two actuators: a coarse stage motorized translator and a fine stage piezoelectric translator. The disturbance delay line is made only by a piezoelectric translator. The coarse stage is used to reach the co-phasing condition starting from a big OPD, i.e. 1 mm. The fine stage works only after the cophasing condition is achieved and is used to control and stabilize the OPD between the two interferometer arms.Co-phasing control system uses only the relative metrology measure and drives the fine stage of the control delay line in a closed loop. The coarse stage delay line is driven in open loop directly by the operator, which reads the necessary displacement to reach the co-phased condition on the absolute metrology monitor. Coarse stage actuator is interfaced with a portable computer by a RS-232 port. software interface realized in NI LabVIEW is used to set all the necessary parameters for programming the actuator driver and for setting the displacement commanded. On a chart is constantly plotted the absolute position of the
actuator. The coarse stage actuator is commanded until a co-phased condition is achieved. In the next figure the experimental interferograms at co-phased condition is shown. Experimental interferograms at co-phased condition. Coarse stage delay line software interface Co-phasing control system hardware is based on a portable computer (P4 2.66 GHz with 2Gb RAM) linked by a fire wire port to a portable NI DAQPad- 6052E port. Even if this type of DAQ board is not a real time board, it is possible to close a digital control loop up to 1 ms control step (clearly not in a hard real time performance!).
National Instruments DAQPad-6052E acquisition board Disturbance delay line actuator is also driven with the same DAQPad-6052E board. A disturbance generation algorithm runs on the same portable computer in parallel to the control algorithm.
Two ADC channels and two DAC channels have been used. The two ADCs channels acquire twosignal coming from the relative metrology electronic used to reconstruct the OPD variations. One DAC drive the piezoelectric driver of the fine coarse of the control delay line, and the other DAC drive the piezoelectric driver of the disturbance delay line. Control algorithm design has been performed following the model-observer based technique Discrete time state equations are directly implemented in a C language algorithm. The control algorithm routine is compiled as a dynamically-linked library (DLL) and interfaced by NI LabVIEW using a Call Library Function Node for exchange data (measures from ADC and commands to DAC) with NI DAQ board. This solution lets to test the control algorithm (written in C language and very close to the final flight version) interfacing it very easily to the laboratory NI DAQ hardware without using flight qualified hardware, saving a lot of time and money. Also the disturbance generation algorithm is implemented in discrete state equations and is written in C++ and compiled in a DLL. In the next figure is shown the control system block diagram:As our system is not a real time system, a very simple software interface has been designed without chart/graphs in order not to load the control algorithm execution. Software interface contains only a set of buttons which are used to start/stop the relative metrology, the disturbance generation algorithm and the control system algorithm and a set of led which indicate the status of the control system. All the important control variables (measures, commands, state variable, etc…) are stored in the computer memory and recorded in binary format on the hard disk at the end of the control session.The test results are very good, the residual OPD has a σ ≈ 9.5 nm and fulfills the requirement. In the next figure are shown the OPD disturbance without control system (left) and the residual OPD with control system (right).In order to improve the co-phasing control system performance, an upgrade of control system hardware is been foreseen. Next development will be the reduction of the control step at 0.1 ms using an NI RT hardware and LabVIEW RT software with a quite adjustment of the control software and control algorithm.

Author Information:
For more information on this Case Study, contact:
F. Musso
ALCATEL ALENIA SPAZIO

Browse All Case Studies »