Vincent finally made the HEPI and ISI behave, but I had a hard time realigning. TMS has a rather large offset in PIT which seems to be caused by HEPI incident.

Anyway, in the end I started over from scratch, centering on ITM using TMS and baffle diodes.

PD1 max (28500 counts) : TMSP=-92313, TMSY=26574

PD2 max (29300 counts): TMSP=-922420, TMSY=11355

PD4 max (29200 counts): TMSP=-54958, TMSY = 11817

I took the mid point of PD1 and PD4 and set TMS offset to (p,Y) = (-73636, 19196).

I used manual alignment followed by tdsdither to obtain this:

ETM: (P,Y)=(-3028.634, 5409.607)

ITM: (P,Y) = (5771.803, 761.433)

REFL_B_PWR went up to about 8100 counts, which is really good (remember, we inserted 10% pick off for WFS, which means that 8100 counts is equivalent of 9000 counts before the change).

However, WFS refuses to work. PIT just goes away even if I reduce the gain considerably. In an attempt to center WFS, one of the picos on the table might have dropped. I haven't looked at the table yet.

I'll leave it WFS-less for now.

Attached is the calibrated spectrum as of now (references are from Aug/25 (blue) and Aug/22 (green), respectively). Somehow 1/f-ish noise between 1 and 20Hz increased, though low frequency noise is good.

The file is here:  /ligo/home/controls/keita.kawabe/OAT_2012/length_example_20120905.xml

The differences between now and then are:

1. Sensor correction is on now.

2. VCO to HEPI relief is off now because Vincent wanted to leave it that way.

3. Alignment is different (HEPI didn't go back exactly to the original position and relatively large realignment of cavity and TMS was necessary).

4. WFS path is different now though WFS is off.

I haven't checked the refcav status except that it is locked.