Thursday, few measurements were performed with the cavity locked. This aLOG presents some results.
In order to keep the cavity locked, Isolation filters on the HEPIs were always engaged, HEPI-BSC6 was yawed by ~180micro radian, damping filters on ISIs, TMS, QUADs and the FM were always engaged. The ISC longitudinal error signal was fed back to BSC6-HEPI (UGF ~1mHz). The simplified control schemes of the BSC-ISI and the HEPI are presented in attachment (SEI_Simplified_Control_Scheme.pdf)

The isolation filters of the 2 ISIs are tuned using the following parameters:
- UGF: 15Hz on all DOFs
- Phase margin > 45deg
- Gain margin>20dB
- Gain peaking <2

The HEPIs are tuned using the following parameters:
- Blend IPS-L4C at 800mHz
- UGF: 10Hz
- Phase margin > 45deg
- Gain margin>20dB
- Gain peaking <2

The measurements were performed using DTT, and then the spectra are extracted and calibrated using Matlab. In order to extract the data from DTT and save them in a .mat file, I used the Matlab and the python scripts called respectively ddt2mlab.m and ddt2mlab.py found in userapps/trunk/cds/common/scripts/dtt2mlab. Due to a bug, I modified (a hack, not a fix) ddt2matlab.py and dtt2matlab.m but I didn’t commit the changes.

The Calibrated spectra presented in attachment were measured in the following configurations:
-          HEPI ON - ISC Y - No Sensor Correction - ISI Damping
-          HEPI ON - ISC Y - Sensor Correction - ISI Damping
-          HEPI ON - ISC Y - Sensor Correction - ISI Damping + Isolation with ST1 L4C blended at 250mHz
-          HEPI ON - ISC Y - Sensor Correction - ISI Damping + Isolation with ST1 L4C blended at 250mHz + ST2 GS13 blended at 100mHz
-          HEPI ON - ISC Y - Sensor Correction - ISI Damping + Isolation with ST1 L4C blended at 100mHz + ST2 GS13 blended at 100mHz
-          HEPI ON - ISC Y - Sensor Correction - ISI Damping + Isolation with ST1 T240 blended at 250mHz + ST2 GS13 blended at 250mHz
-          HEPI ON - ISC Y - Sensor Correction - ISI Damping + Isolation with ST1 T240 blended at 100mHz + ST2 GS13 blended at 250mHz

The plots show spectra of the length of the cavity (LHO_OAT_ALS-Y_ARM_LONG_IN1_DQ_2012_09_06.pdf), the motion of stage 1 (LHO_OAT_ISI-ETMY_ST1_BLND_Y_T240_CUR_IN1_DQ_2012_09_06.pdf) and stage 2 (LHO_OAT_ISI-ETMY_ST2_BLND_Y_GS13_CUR_IN1_DQ_2012_09_06.pdf) of ISI-BSC6 (ETMY). The mystery noise (fiber?) mentioned in 4128 slightly disturbed the measurements; it is visible on some spectra (on the black curve of LHO_OAT_ALS-Y_ARM_LONG_IN1_DQ_2012_09_06.pdf around 2 Hz for instance).

On the spectra of the cavity, the effect of the sensor correction on the HEPIs is noticeable (blue vs green). In both cases, the ISIs are only damped. However, there is a significant amplification below 100mHz (corner frequency of the STS-2 (on the ground) high pass filters at 40mHz). Above 100mHz, there is a reduction by a factor of 3-4 visible up to 2-3Hz.
Once the isolation filters are engaged on stage 1 only (blend at 250mHz on the L4C – in red), the attenuation is important above 250mHz and the amplification is still visible below 100mHz (should be amplified by the stage 1 CPS low blend filters).
Next, stage 2 isolation filters were also engaged and blend frequencies were lowered down to 100mHz and finally the T240s were introduced in the stage 1 super sensor (CPS + T240 + L4C). The extra isolation (low blend + T240s + 2 stages controlled) provided by the ISIs is not visible on the cavity above 200mHz (reaching the noise floor of “something”). Only the amplification below 100mHz is visible due to the use of aggressive blend filters (100mHz) and the 2 stages of isolation. The largest amplification is obtained when the 2 stages are controlled and the position sensors are blended with the seismometers at 100mHz.

The isolation provided by the ISIs is presented in figures (stage 1 - LHO_OAT_ISI-ETMY_ST1_BLND_Y_T240_CUR_IN1_DQ_2012_09_06.pdf) and (stage 2 - LHO_OAT_ISI-ETMY_ST2_BLND_Y_GS13_CUR_IN1_DQ_2012_09_06.pdf). The best isolation is obtained when the T240s are introduced in the stage 1 super sensor with a blend frequency of 100mHz. Actually, the CPSs are blended with the T240s at 100mHz and the T240s with L4Cs at 2Hz (cf scheme control).

Comments on the amplification at low frequency (peak at 60mHz)?
 At 60 mHz, when the ISIs are not controlled, the ISIs absolute motions are about 1 micrometer and the relative motion between the 2 ISIs is 100nm. A transfer function from the STS-2 at the end station to the STS-2 at the corner station showed that the LVEA and the end station are moving in phase around 100mHz (a 5-6 degrees phase would explain the 100nm of relative motion between the ISIs).
Once controlled, the absolute motion of the ISIs is amplified by 10 but what is the phase between the 2 ISIs? I tried to measure transfer functions from the T240s in BSC6 to the T240s in BSC8 but the coherence is close to zero.
Under control, the relative motion between the 2 ISIs is 10^4nm (x100 amplification). A phase of 60 degrees between the ISIs would explain that. But, it would be surprising to see this large phase since the filters used on both ISIs are quasi identical.

At low frequency, the platforms are locked to the ground using the position sensors. But if the so called vertical position sensors are not all “perfectly” aligned with the vertical, a translation of the platform in the horizontal plan will create some tilt. Under control, if the ISIs are moving in phase in the Y direction but one ISI is tilting in +Rx while the other one is tilting –Rx, the variation of the length between the test masses hanging points will be increased.

The tilt correction has not been implemented on the ISIs. Few weeks ago, I quickly tried but I didn’t see any significant improvements. It may be worth trying again.