[Miyoki, Nakano, Miyazaki, Enomoto]

We switched off all the air flow related things we can stop and checked the control signal spectrum of Gr PDHX. And fount it improved!

We switched off KOACH for PSL room, air temerature controllers for PSL, an air compressor between PR2/3, and FFU clean filters for PR-MCF area and BS area.
Here is the comparison of the control signals with various conditions.
Black trace corresponds to all-on condition, red one to all-off condition, and cyan one corresponds to FFU for the condition where PR-MCF area was on. Blue one corresponds to the error signal of IMC loop taken simaltaneously with cyan one.
Improvment above ~200Hz comes from switching off the KOACH, and improvement below 100Hz comes from switching off air temerature controllers for PSL (and partially from FFUs).
It turned out that operating one of the air temperature controllers does not increase the noise significantly.

And then, we closed ALS CARM loop and saw if we can keep the IR resonance or not.
Here shows the transmission of IR when the loop was closed. This became significantly stable.
The condition for the following figure: one air temperature controller was on and the air duct was half-opened.

Note:
-- From the PDHX control signal spectra, it apears that FFUs do not have much influence on ALS, but they do. By switching off FFUs, the IR tranmission was calmed down very much. This implies that fiber phase noise induced by the vibration of the cable rack limits the performanc of ALS.
-- How air temperature controllers induce noise on ALS is not so clear, but one possibility is through the vibration of the wall of PSL room, similarly to the above noise path.
-- We still see the frequency fluctuation in IR transmission, at several tens of Hz and ~0.2-0.5 Hz. Still do not know where they come from. Not enough gain in microseismic region? some scattering? just another fiber phase noise?