Presentation on theme: "10-07-08ILIAS WG1 meeting - Birmingham1 Cryogenic Trap Vs Brewster window & noise coupling through windows J. Marque."— Presentation transcript:
ILIAS WG1 meeting - Birmingham1 Cryogenic Trap Vs Brewster window & noise coupling through windows J. Marque
ILIAS WG1 meeting - Birmingham2 Motivations The Brewster window was supposed to be the main coupling path to the dark fringe for acoustic and seismic noises in the central building. Below 100Hz Above 100Hz AC off test
ILIAS WG1 meeting - Birmingham4 Cryo Trap benefit (2) Highly improved stationarity respect to alignment fluctuations / bad weather Before After Before After Before After
ILIAS WG1 meeting - Birmingham5 Cryo Trap isolation performances (1) Cryo Trap Residual Gas Analyser SR valve DT valve -196 degrees (at this temperature hydrocarbons will condensate when hitting the tube)
ILIAS WG1 meeting - Birmingham6 Cryo Trap isolation performances (2) DT contamination level DT RGA (blue) vs Cryo Trap RGA (black) larger by > 10 times before DT opening after DT opening CryoTrap RGA: no increase of contamination on SR/BS side after opening to DT
ILIAS WG1 meeting - Birmingham7 Cryo Trap isolation performances (3) after DT opening after SR/BS opening CryoTrap RGA: no significant increase of contamination on SR/BS side after opening to SR
ILIAS WG1 meeting - Birmingham8 Cryo Trap maintenance requirements LN2 consumption : 2 refill / week (Friday-Tuesday), tolerating a 3-4°C of warmup of the upper side LN2 delivery every 2 weeks on tuesday: small truck at central building, lift needed, time of arrival difficult to be programmed regeneration (needed every ? Months) maybe take more than 1 day. To be studied LN2 supply for V+ to be discussed, solution possibly implemented during the summer stop
ILIAS WG1 meeting - Birmingham9 What was the noise coupling mechanism? Would have been interesting to understand exactly what was the coupling mechanism not to repeat this mistake in the future… We have still windows: input/output tank windows, Injection Brewster windows. - Elasto-optic effect (seismic/acoustic produce a variable stress on the window. The elasto-optic effect translates that into index variation, which modulate the phase of the beam)? - Back-scattered light (some light will recombine to the main beam phase modulated by the scattering object motion)? - Clipping / scattering on tubes can be excluded (many diaphragm baffles have been installed before and after the Brewster without any improvement) …
ILIAS WG1 meeting - Birmingham10 Concrete example: the detection output window (1) After first relock with the cryo trap: noticed coherence with detection tower accelerometer attached close to the output window. Cryogenic trap requires the detection turbo pump to be always on (was always off with the Brewster window) Config: Cryotrap and turbopump on Config: Brewster and turbopump off
ILIAS WG1 meeting - Birmingham11 Detection turbo pump problem fixed with bellow installation (2) Installation of a bellow between detection tower and turbo pump. Turbo pump clipped to the platform.
ILIAS WG1 meeting - Birmingham12 Detection turbo pump problem fixed with bellow installation (3) Turbo on Turbo off Comparison between: SR+DT turbo pumps off SR on + DT off SR on + DT on with bellow Remaining coherence: there is still something!!!!
ILIAS WG1 meeting - Birmingham13 A tentative projection of the noise coupling through the detection output window (1) Assumptions: - the turbo pump noise path to dark fringe was all through the DET window, and not for example through the detection bench. This assumption is supported by the fact that the seismic noise produced by the pump is larger on the flange than on the bench. - the Em_SETODE01 accelerometer is actually measuring the vibration of the DET window. This sensor is attached to the tower flange very close to the DET window, and perpendicular to it. The coherence is however not so good: it is between 0.1 and the noise coupling is direct, not through upconversion (this is supported by the fact that there is some coherence).
ILIAS WG1 meeting - Birmingham14 A tentative projection of the noise coupling through the detection output window (2) Transfer function, in units of strain/m, between the sensitivity and the seismometer. Indicated in red are the coherent frequencies (coherence > 0.1). That is were the measurement makes sense, elsewere it is just an upper limit of the real TF. Projection of the detection flange seismic noise. The upper plot shows how the projection reproduces the dark fringe noise structures at Hz and 350Hz before the DET turbo pump was damped. The bottom plot is the projection after pump was damped. In both cases the red line is an upper limit of the noise. Good reason for putting the diodes under vacuum
ILIAS WG1 meeting - Birmingham15 Noise coupling through the end towers output windows We have no measurement… But since we know that here the problem can only be backreflected or backscattered light AND since we have to replace this window (damaged by the hublot implosion), it is time to speak about requirements. 1)Backreflected light: the mechanical mount of the window will be changed so that the window will be tilted by a few degrees. The window will be AR-coated in LMA (<100ppm). 2)Backscattered light: here are the requirements for the polishing and corresponding amount of scattered light. -Cosmetics 20/10 => TIS=150ppm -Roughness: TIS=35ppm Theoretical (Parks, 1980), on 100mm diameter
ILIAS WG1 meeting - Birmingham16 Conclusion The cryotrap fixed the Brewster link noise coupling. But, so far, we are not sure about the coupling mechanism. - The detection output window is the new dominating coupling path => diodes into vacuum. - The end towers output window will be carefully analysed. - What about the injection Brewster? The input window?