July 2001 J-P. Berger Summary of the June-July 2001 run from the IO point of view ----------------------------------------------------------- INITIAL GOAL: The goal of this run was to install a permanent optical interface on the former infrared table allowing to send light coming from the three telescopes into an integrated optics beam combiner, to record fringes on three baselines. Since the pixel readout mode was not available and therefore no fringe recording was possible the main interest in the experiment was to test separately all the interface elements. INTERFACE DESCRIPTION AND TESTS: A full description of the optical interface will be found in the following web site: http://cfa-www.harvard.edu/~jberger a) scanning stage: Description: ---------- This part is common to both free space and IO beam combiner. It is made of three dichroics mounts located on the same place. Two of the dicroics can be scanned back and forth with piezos. In the absence of dichroics they were replaced by mirrors. Tests: ------ We setup a Michelson classical interferometer to test the scanning properties. One of the interferometer mirrors was one of the 2 scanning mirrors. We use a He-Ne laser and visualise at the oscilloscope the laser interference fringes produced by successively each of the two scanning mirrors and detected by a photodiode. The scanning ramp was a triangular waveform of amplitude 0 to 10 V. Both piezo showed the same properties. We found no evidence of tilt throughout the scanning process. The laser fringes show no change in contrast down to a 5% contrast accuracy along the piezo range. We explored ramp frequency parameter space from 1Hz to 10 Hz. We found that the fringe contrast decreased with freq which is not attributable to a tilt (the tilt was adjusted to maximise the contrast for each frequency). This could be attributable to the detector cutoff freq (not probable) or the oscilloscope itself. We found that the fringe envelope showed damped oscillating pattern with a relaxation time going from 10% to 25% of the whole scanning ramp, the longer time being obtained at 10 Hz. This oscillation was obtained at each turn of the triangular ramp. A part fron this effect the piezo+mounts behavior can be considered satisfactory To be done: ---------- The oscillating effect needs further analysis in order to reduce it (mechanical mount). The dicroics plates will eventually be located at one corner of the IR table. JPB thinks there will be not enough room for them because of the 20 deg angle constraint and the telescopes mounts. Will it be sufficient to just cut the plate ? b) Focusing module Description: ----------- It consisted of three plates containing the oap mounts and 4 axis fiber positioners. xy,z(focus),theta(fiber axis) lateral xy axis are motorized by servoed piezo. Tests: ------ The OAP mounts did work OK. Put 3 pieces of shim under each pushing point to avoid bending. There were very thin scratches all over the surfaces of the 3 parab. Should disappear with the new coating. The image of the fiber at the focus show nearly no aberration going through focus show a nearly circulary (slighly elliptic) symmetric pattern. During star observation and coupling optmisation the piezo resolution proved to be sensitive enough. One DC offset knob turn ~2 pixels on the star tracker. To be done: ---------- Silver protected coating Software control of the piezos offsets. Test fiber theta mounts. c) Integrated optics beam combiner. Description: ----------- Its a pairwise beam combiner manufactured at LETI (silica etching technology). 3 inputs 9 outputs` : 6 interferometric and 3 photometric. Tests: ----- The BC was found broken in is protective box. The break occured at the interface between the fiber vgroove and the IOBC. The packaging was obviously not adapted. Patches of glue remained on both surfaces. We cleaned up both surfaces by polishing with 1 micron and 0.3 micron sanding paper and alcool. We got rid of glue patches. We aligned dynamically fiber vgroove and IO. To do that we setup 2 different mounts one holding the IO BC (one free degree: angle around optical axis). the other the vgroobe (5 axis). The interest of this alignment is secondary since we hope it will be the last time we need to do that on the mountain. Because we lack an axis to rotate the vgroove around the optical axis the global alignment of the fiber+output+pixel column was tricky. We did not want to loose time on shiming issues.Therefore the simultaneous optimisation of three ways was far from satisfactory. We noticed that the three vgroove fibers were not aligned, the central one seems off. The use of liquid matching index improved by a factor of 60% the coupling efficiency between fiber and IO. We observed the IOBC output with vega and Mu Cep with one telescope using and idl written visualisation tool: Nice interface program alinging integrated optic component on Quadrant -- ~iota/Idlstuff/EP_Scan/customDAQ.script. We succeeded in getting counts on both stars. We don't know exactly what was the integration time in this quadrant readout mode roughly 0.2s. Mu Cep (July 3) Max counts on one interferometric output: 30000 Noise level~100 Vega (July 3) Max counts on one interferometric output: 2600 Noise level~100 To be done: ---------- Connect a new beam combiner with fibers and add a permanent protective package. d) Aligment module. Description ----------- Four pellicular beam splitters+ telescope mount + fibered collimator+corner cube. Tests: ----- Concept OK but optical surface quality of pelliclar BS really bad. This prevented us from coupling white light in the three fiber simultaneously. In its present test this interface doesn't allow to record internal white light fringes. It was only useful with the visible laser to light up simultaneously the three inputs and look at them on the visible camera. To be done: ---------- Find better beam splitter and redesign the interface consequently.