We measured power spectrum of signal from tilt sensor in a variety of situations made by operating fans' switch, cover(made of aluminum foil ; just putting on the pylon) and oplev laser in stationary. And our data is the average of 20 times. The results are attached.
Hi tilt-sensor folks, as a reference I put one of links to oplev data here. Here the suspended PR3 mirror was monitored by several sensors including an oplev; the resonance peak of the suspension could be seen with the oplev (with an appropriate wind shield).
Do you think the tin foil cover is not enough (compared with your appropriate wind shied) and that's why it's still noisy with fans off and covers on? We are also suspecting some upconverted scattered lights. However the main MCE merror is the only one suspicious scattering object as nothing else shouldn't be moving very much much.
Well, I'm sorry but I didn't have a look how the foil-covers were covering the optical setup of the tilt sensor, so I cannot say whether it is enough or not....
If the tilt sensor's output would be dominated by some scattering of light, the spectra of the oplev staying with the tilt sensor would show different shapes even for the same mirror (MCe?), but I think both spectra (of the tilt sensor and the oplev) basically show the same tendency (is my understanding correct?); if this is the case, the noise might not be due to scattering light. The shapes of the spectra (below 30Hz??) still seem similar to the ones due to air turbulance.