[Park, Kokeyama, Kirii, Cho]

Last week we found the oplev got much higher noise compared with the configuration change on August 15th. Followings are what we tried and fixed:

- Checking the oplev hardware in the tunnel, we found that the oplev beam is hitting on a very very edge of the MCE mirror. Indeed, oplev SUM channel is decreased from 21000 (8/14 JST) to 16000 (current. see, oplevsum.png). Hitting on the edge is very bad (i.e. OL must have very big angle to length coupling and calibration factors to urad must be off for some amount) but there seems no option for the oplev beam path to live with the tilt sensor.

- We also noticed that the tilt sensor laser source was clipped a bit by the optical fiber of the oplev. The fiber was fixed by a clamp softly and not not clipping the tiltsensor beam.

- Reflected and scattered light by the oplev QPD was going back to the viewport. We blocked the junk beam.

- To avold the air fluctuation, we covered the pylons of the laser sources and detection pylons with tin foils.

- Dark noise (and also electric) noises were confirmed not to limit any signals (see, darktest0822.png).

- We found that phase unrapping c code by Junegyu accumurates the phase forever. This time we reset it by restarting the model, however, we must fix the code and add a "reset" botton on the medm screen.

The resulting data is spctrm0822_PIT (YAW) .png. In low frequencies, oplev and tilt sensor disagree for about the factor of 5. Oplev signals are still very bad compared with Aug 14th, when was before we change the optical configuration of the OL laser and QPDs. I think the current oplev configuration is too bad. And we have no idea what is wrong with the readings of tilt sensor. Tilt sensor channels must be giving the absolute phase in rad after doing arctangent ...