Automatic Alignment using the Anderson Technique A. Freise European Gravitational Observatory

1. Automatic Alignment using the Anderson Technique • A. Freise • European Gravitational Observatory • Roma21.10.2004

2. Overview Output Mode-Cleaner • Linear alignment • Drift control • Non-linear alignment • Simulation • Procedure/Documentation • Automation Suspended bench External bench

3. Linear Alignment: Status Output Mode-Cleaner B8 • Linear alignment implemented for North arm, West arm andthe recombined Michelson, using B7 and B8 • Performs well for full power or reduced power (10%) B7 Suspended bench External bench

4. Autoalignment: Why ? • Superimpose beam axes • Maximize light power • Stabilze optical gain • Center beam spots on mirrors • Minimize angular to longitudinal noise coupling

5. Autoalignment: How ? • Differrential wavefront sensing (analog feedback for 14 DOF in GEO) • Spot position sensing (digital feedback for 20 DOF in GEO)

6. The VIRGO Interferometer 2 Perot Fabry cavities W N Recycling mirror EOM Injection Bench

7. ‚Linear Alignment‘ for VIRGO linear alignment : angular motion of 5 mirrors to be controlled (DC – 4 Hz)

8. Modulation-Demodulation For obtaining control signals a modulation-demodulation technique is used. Only one modulation frequency is applied to generate all signals for longitudinal and angular control of the main interferometer. 6.26 MHz

9. Resonance Condition TEM00 Upper Sideband Carrier Lower Sideband

10. Resonance Condition Upper Sideband Carrier Lower Sideband TEM01

11. Cavity Alignment The Anderson technique uses signals in transmission of a cavity. The detectors are positioned in : Near field Far field

12. Cavity Alignment The Anderson technique uses signals in transmission of a cavity. The detectors are positioned in : Near field Sensitive to translation of the mode Far field

13. Cavity Alignment The Anderson technique uses signals in transmission of a cavity. The detectors are positioned in : Near field Sensitive to tilt of the mode Far field

14. Detection

15. Detection • In each of four outport ports • we can set: • two Gouy phases • two (four) demodulation phases • to get 4x4 output signals for • each direction (horizontal/vertical)

16. Detection For tuning the telescopes one can move L2, L3, L4a and L4b. The most critical adjustment is required for L2.

17. Tuning Telescopes

18. Control Matrix In total: 8 Gouyphases have to be tuned, 16 demodulation phases to be set. This yields 32 signals to control 10 degrees of freedom(5 horizontal, 5 vertical). Control topology (phases+control matrix) has been designed by G. Giordano. The optical matrix has to be measured to generate two 5x16 control matrices using a 2 reconstruction method.

19. Example Matrix (16x5)

20. Signal Amplitudes

21. Alignment Control DC: beam positions are defined by reference marks, spot position control, below 0.1 Hz around the resonance frequencies of the suspension pendulums the beam follows the input beam from the laser bench, differential wave-front sensing, 0.1 Hz to 10 Hz no active control at the expected signal frequencies, the two mode cleaners suppress geometry fluctuations by ~106

22. The GEO 600 Detector +2 for MI differential mode +2 for signal recycling 16 spot position control 4 degrees of freedom for MC 1 +4 for MC 2 +4 for MI common mode + 32 = 48 differential wave-front sensing

23. Signal Amplitudes in 2D

24. Zero Crossings

25. Angular Fluctuation • Residual fluctuations:~ 1 nrad @ 10 Hz~ <1urad RMS

26. Filter design • open loop transfer function for NI/NE tx. unity gain 3.2 Hz

27. The Suspension Control • Main mirrors are suspended for seismic isolation. Active control is necessary to keep the mirrors at their operating point: • inertial damping • local damping • local control, i.e. steering of the mirrors • Bandwidth ~5 Hz, positioning of the mirror to ~1 mrad and <1 mm Good performance for operating the interferometer but more precise controls are necessary to reach the expected sensitivity of the instrument.

28. Feedback Feedback is applied to the Marionette via the four coil-magnet actuators used alsofor the local control.

29. Current Status Output Mode-Cleaner • Interferometer currently used in recombined mode (Recycling mirror is misaligned) • North and West arm cavities are automatically aligned (to the beam) since: • North arm: December 2003 • West arm: May 2004 • Longest continuous lock >32h • Beam drift correction not yet implemented Suspended bench External bench

30. Cavity Power AA Off AA turned ON

31. Angular Fluctuation AA ONAA OFF • From Local to Global control • Bandwidth ~4 Hz

32. Angular Fluctuation • Residual fluctuations:~ 1 nrad @ 10 Hz~ <1urad RMS • Limited by: • input beam jitter • resonance peaks of the main suspensions (e.g. 0.6 Hz)

33. Conclusion Output Mode-Cleaner • First implementation of the Anderson technique on a large scale interferometer • Both arms of the interferometer are automatically aligned: • Local controls can be switched OFF • The angular mirror motions are reduced and the power fluctuations of the arm cavities minimized • Facilitate the recombined lock acquisition • Unity gain frequency around 4Hz • 32 hours continuous lock of the interferometer with automatic alignment control • Next steps • Beam drifts correction • Recycling mirror automatic alignment

34. End

35. Global Control Output Mode-Cleaner • 8 quadrant diodes yield 32 signals • Signals are linearised by the DC power on the quadrant • A static matrix is used to create 10 signals for angular control of the mirrors • Unity gain bandwidths is 3 – 5 Hz • Automatic alignment allows switch off the Local controls