OMC Length control loop was evaluated to estimate the out-of-loop stabilization displacement level.
The out-of-loop stability of the OMC is around 10^-14~10^-13m/rtHz level (Attachment 3 blue curve below 10Hz).
Between 10Hz to 300Hz, the contamination of the cavity length displacement is expected (Attachment 3 black curve).
This is due to the sensor noise injected to the cavity by the dither control loop itself. 

The data was taken from the lock in the evening of Apr 15 at 15W lock.

Servo filter modeling

The transfer function from H1:OMC-LSC_I_OUT_DQ to H1:OMC-LSC_SERVO_OUT_DQ was measured and modelled
based on the ZPK model of the filter. (Attachment 1)

Openloop TF modeling

The OLTF model is based on a swept-sine loop TF measurement done on Apr 4 at DRMI with sideband locked OMC.
As the DCPD output is normalized with the DC value, the optical gain after the normalization is believed to be constant.

The actuator frequency response is indeed not simple, but we already evaluted it before (LHO ALOG 16089).

In the end, the openloop transfer function was successfully modelled. (Attachment 2)
Note that the actual open loop transfer function on Apr 15 was a factor of 3 lower than the one in this plot.
(The servo gain was 30 on Apr 4, and was 10 on Apr 15.)

This gives us the optical gain estimate of 6.0x10⁸ 1/m at H1:OMC-LSC_I_OUT_DQ.

...In fact this number is about 5 times smaller than the theoretical prediction from the dither amplitdue (0.87pm_pk@3300Hz)
and the cavity finesse of ~400. I still don't have good excuse for this discrepancy.

Free-running and stablized cavity length displacement

The free-running motion of the OMC cavity was estimated from the optical gain, actuator response, and openloop transfer functions.
It is shown as the red curve in Attachment 3. The blue curve is the calibrated in-loop displacement sensed by the dither length sensing.

The gray curve indicates the cavity length noise measured with PDH in January (LHO ALOG 16089).

Assuming the red curve is dominated by the sensor noise (and that is the guess), the black curve is the noise injection to the OMC cavity
due to the servo control. This means that the out-of-loop stability of the OMC is represented by the blue curve up to 10Hz,
the black curve up to 300Hz, and the gray curve above 300Hz.

The sky blue curve represents the shot noise level estimated from the shot noise level of the 33mA and optical gain 6.0x10⁸ 1/m.
(i.e. Sqrt(2 e / Idc) * 3000 / Sqrt(2) / 6.0x10⁸ = 1.1 x 10^-14 m/rtHz, 3000 is the LO amplitude for demodulation,
1/Sqrt(2) is the product of the demodulation (1/2) and double-sided to one-sided conversion (xSqrt(2)) )

The information we can obtain from this measurement is:

1) Below 4Hz, the dither error signal sees the cavity motion. (Compraison between red and gray)
2) At least above 10Hz, the sensing noise is dominating the error signal. (Comparison between red and gray)
3) The cavity noise have some mechanical peaks at 0.45Hz and 2.8Hz. (Comparison between red and gray)
    Where are they coming from? Is this due to OMC SUS actuation for OMC ASC?
4) The cavity motion is contaminated by the control above 10Hz upto 300Hz. (Comparison between black and gray).

You may wonder:
a) if the bump at 30~40Hz comes from the sensor noise, or coming from the servo bump or something else.
=> My guess is that this is the sensor noise. If you look at the DCPD spectrum (Attachment 4), the structure around the dither
frequency is asymmetric. This seems indicating that there is a bum ~3330Hz and a part of the bump is suppressed by the servo
while it is adding more noise to the lower side of the dither frequency.
=> I'd recomend to shift the dither frequency where there is no structure in the DCPD spectrum with in ~+/-100Hz of the dither freq.

b) what are the strong peaks above 100Hz. They are the down converted violin modes from ~3000 and ~3500Hz.
This is also visible in the DCPD spectrum (Attachment 4). Note that this strong harmonics of the violin modes are partly coming from
the saturation of the DCPD during this lock.
=> I'd recomend to use stronger roll-off or band-stop. Also the modulation frequency should probably be selected to push the violin
peaks in a small freq region in the dither error signal.

c) if it is ok to have the contamination if the cavity motion due to the dither locking.
=> We probably need to tune the servo filter to minimize the contamination to keep the requirement of 3x10^-16 m/rtHz.
The design depends on how the noise bump at 35Hz is removed by shifting the modulaiton frequency.

Elli, Kiwamu, (under WP 5166)

In response to Evan's report from the last night (alog 18020) about the ITM cameras not reproducing good recycling gain, we took a close look at the ITM digital cameras. As a part of the activity, we removed the lexan plates from the following cameras in order to improve quality of the images:

• ITMX infrared spool camera (VP2)
• ITMX green spool camera (VP1)
• ITMY infrared spool camera (VP5)
• ITMY green spool camera (VP6)

This is going to be a permanent configuration. The removal of the lexan got rid of unwanted stripes that had been seen on the ITMX green camera. The attached files show the ITMX green camera before and after the lexan removal. Hopefully this will help the reproducability of the initial alignment. Note that the oval shape shown in the attached pictures are due to the software aperture masking. The ITMY camera did not have stripes from the beginning and therefore no drastic change between before and after. We did not check the quality of the infrared cameras yet as they had not been an issue. We are still in the middle of figuring out optimum combination of the analog gain and exposure time, but we handed the interferometer to the evening commissioners.

Prior to the removal, John came with us to the LVEA and we went through all four cameras in order to see if they are mounted properly such that they don't accidentally hit the viewport within the movable range of the mount. Partly because of this test, we had to readjust the camera angles afterwards which then required us to renew the ALS ASC camera reference positions. We started the alignment process from TMS on both arms using the baffle PDs and subsequently optimized the ITM and ETM angles by maximizing the green transmission. Then we updated the camera positions. 

After Tuesday PZT mirror swap, the ISS photodiodes are misaligned, see attached trend (all ISS signal vertical scale is 0-2000). All diodes receive less than half the power they used to. Some even less than that.

07:00 Cris and Karen into LVEA

08:00 OIB set to UNDISTURBED

08:45 OIB set to Commissioning

08:46 begin alignment/locking

09:02 J Bartlett out to the LVEA to hang sensors for 3IFO

09:15 Bartlett out of LVEA

09:40 Fil and Andres out of LVEA

11:00 Operators meeting

12:00 Operators meeting over

14:02 Kyle to EX

15:15 Kyle back from EX

I did some initial alignment today. Locking happened up to the Resonance state. I never experienced the oscillation mentioned in Sheila's aLog (comments to 18020). IFO broke lock within 3 minutes of resonance 2 - 3 times. It never got past that then commissioners took over.

Had "tinkered" with out-building IP controllers -> As found, Y-mid was 5000V and 7000V -> changed current at which point one of the channels stepped from 5000v to 3000V such that now IP9 is 3000V and 5000V

Aidan, Elli, Nutsinee

Following up Elli's alog 17993

I took advantage of a ~9:30 lock loss this morning to rotate the CO2X rotation stage around (the operator was informed). It took longer than I expected so I accidentally worked through as the Guardian was building up IR power (if you see something weird during 9:30-9:50 that could have been me. Sorry...)

Procedure: I requested the rotation stage to search for home, then requested a power of 0.212 W (nominal). Then repeatedly go to minimum and request the same power again to see if the rotation stage accumulate some error in angles. Eventually, I searched for home again a couple times then brought the power back to nominal.

Result: I find the result quite puzzling and need some times to think about it, but it seems like after search for home the difference between the requested power and the output power grows significantly. However, after a few adjustment I managed to get the requested nominal power and the output power to agree within 1% and that's where it was left this morning.

I had a look at the SRCL coupling to DARM during last night lock. Indeed it is non stationary, and the transfer function changes shape significantly. It appears, as already seen, that there is a low frequency component which is more or less constant, plus a high frequency component with a quickly varying gain. See the first plot for an animation of the TF from SRCL output to (calibrated) DARM.

I analyzed the coupling at 40 Hz and at 300 Hz. For the low frequency, the fluctuations are mainly in the transfer function gain, which is on average about 3e-11 (in some units) and fluctuates between 2 and 8e-11. For the high frequency, it seems that the phase as well as the amplitude changes. So I analyzed the real and imaginary parts of the TF separately.

The attached plots show that the coupling fluctuations are very well correlated with angular motions. In particular H1:ASC-AS_A_RF36_I_PIT_OUT_DQ is the winner. This signal has the same content as H1:ASC-AS_B_RF36_I_PIT_OUT_DQ.

So my conclusions is that the non stationarity is still dominated by angular motions in the SRC. Indeed, we didn't switch on any ASC boosts on SR1/SR2. We have to improve SRC angular stability. In particular SRC1_PITCH

Scott L. Ed P. Chris S.

This report will cover the following dates: 4/21 & 4/22.

On Tuesday the crew cleaned 46 meters ending 10 meters north of HNW-4-008. Test results posted here.

Wednesday was spent removing and re-hanging lights, applying bleach/water solution, vacuuming support tubes and safety meeting.

Sheila, Gabriele, Evan

Tonight we spent some more time refining the locking sequence with good recycling gain. Some lessons learned:

• It seems that saving the green spot positions on the ITM cameras is not enough to consistently bring us to a good recycling gain. Instead, it seems that the spot position on the POP camera and POP QPDs is a better indicator; a high spot seems to give a mediocre recycling gain. To reduce the amount of ITM steering required in full lock, as part of the initial alignment (with IR locked to the X arm and with WFS closed around IM4 and PR2) we moved ITMX in pitch by about 2 µrad to bring the beam lower on POP A. This brought us to an okay recycling gain (about 34 W/W) in full lock. It remains to be seen how repeatable this is.
• In addition to the bounce and roll mode work this afternoon, we spent some time damping violin modes on ITMY, since they were badly rung up. In particular, 501.750 Hz is damped with a gain of +200 and 0°, and 501.811 Hz is damped with a gain of -200 and 0°.
• We found that the steering we did with the ITMs tended to cause an ~0.5 Hz angular instability (seen in the ITM oplev pitch signals) when we powered up from 3 W to 10 W. Since ITM pitch is not servoed, we suspect this is radiation pressure coupling caused by beam miscentering.
• Increasing the power to 10 W tends to spoil whatever ITM alignment work is done at 3 W (we see the transmitted arm powers drop). We again suspect radiation torque.
• We were able to engage the new LSC boosts (DARM, MICH, SRCL) without issue.
• On dc readout, we injected broadband noise into SRCL. 10:55:15 to 10:59:15 UTC is done with feedforward off, and 11:04:00 to 11:09:00 UTC is done with feedforward on. In both cases it seems the coupling is still nonstationary.

As reported in the previous alog (17990), we noticed that the IM4 trans has shifted by -0.7 counts horizontally after the pzt move (alog 17976) yesterday. I think that we have introduced a horizontal translation in the beam on the PSL mirrors by about 2 mm during the pzt move which resulted in the different IMC output beam alignment. I corrected the alignment of the IMC output beam, by moving IM2 and IM4 this morning.

(Estimation of the shift)

I did an A2L measurement (see Jax's alog 6676 for more details) on MC1 and MC3 to measure the amount of off-centerings. Since we have an active control for the MC2 spot position (DOF3), I did not measure it assuming the in-vac MC2 trans QPD has not moved. The table below summarizes the measurement from today with ones from Dec-2013 for comparison:

Since we know that the vertical position on IM4 trans did not change this time, we ignore them for now. Apparently the horizontal position shifted by 1.9 mm and -2.4 mm on MC1 and MC3 respectively. Note that the MC1 and 3 have a beam radius of 2.12 mm (see T0900386-v7) and therefore the shift is as big as the beam size. If we believe these numbers, the beam must have shifted as depicted in the below cartoon:

Assuming that the distance from a mirror at the bottom of PSL periscope to MC1 is roughly 8 m, the translation we have introduced is about 2 mm toward the West, which I think does not sound too crazy.

If we want, we can correct this by steering a combination of some mirrors in the PSL so as to compensate the translation.

(Doublecheck with MC suspension biases)

In addition to the spot position measurement, I also checked the suspension bias sliders for double check. According to 2 days trend of the suspension bias values, it seems that the ASC servos introduced 35 urads differential yaw on MC1 and MC3 while MC2 did not move much in yaw after the pzt move. I attach a screen shot of the trend. Though, I have no idea what was happening in the vertical directions which also show some variation. Based on a geometrical arrangement of the IMC alignment (for example, J.opt.13 055504 ), one can write the amount of mis-centering on MC1 and OMC3 due to differental angles as

x_mc1 = -x_mc3 = 2 * sqrt(2) * L * theta_mc1

Substituting theta_mc1 = 35 urad, one can obtain x_mc1 = 1.6 mm for MC1 (and -1.6 mm for MC3). So this is consistent with the spot position measurement and explains large fraction of it.

(Moving IM2 and IM4)

I decided to correct the output beam of IMC by moving some combination of IMs in order to keep the same PRM spot position. Since I did not like moving IM1 as it is far from the input Faraday, I decided to move IM2 to recover the spot position on IM4 trans. Then I corrected the angle of the beam by steering IM4. Here is a table summarizing how much I moved the bias sliders.

The direction of the move on IM2 is consistent with the above arguments. Fortunately, the move on IM2 and IM4 were such that they reduced the absolute numbers.

Added 1233 channels
Removed 416 channels

74 channels remain unmonitored (list attached)

8:53 Travis to LVEA

9:00 Travis back

9:59 Metal recycling container going on the truck

10:06 The truck left

10:08 Karen to Mid Y

10:57 Karen leaving Mid Y

14:13 Jeff B. to 3 IFO storage area in the LVEA

13:24 Jeff B. back

After the PSL periscope PZT swap, the periscope bottom mirror is a different high reflector.  We are using the transmission through that optic to monitor the input power into the interferometer; since the transmission of the new optic seems to be a little higher than the old one our calibration of the power into the interferometer has been off since tuesday.  We used to get a maximum power input of 22.7 Watts, after the swap our monitor reads a maximum power of something like 23.7 W.  I adjusted the responsivity of the periscope PD, it used to be set to 0.35 and is now 0.311.  

Next time someone is in the PSL it would be good to check this calibration using a power meter.  

Detchar would like to request that the HWS cameras in the center building be turned off for a few minutes at a known time. We're trying to track down some glitches in PEM and ISI sensors that happen every second, and Robert suspects the HWS. Just a few minutes with them off, and then on again, would be fine; we don't need the IFO to be in any particular state, as long as the ISIs are running fine. We would need the precise times (UTC or GPS preferred), as the channels that record the camera state don't seem trustworthy (alog).

Andy, Duncan

When looking for times when the HWS camera was on or off, I found that the minute trends indicated that it was off on Apr 18 6:30 UTC for ~27 minutes. But the second trends indicate that it was turned off 20 minutes later than that (and back on at the same time). The raw data (sampled at 16 Hz) indicates that the camera was never turned off.

This was originally found using data over NDS2, but Duncan has confirmed by using lalframe to read the frames directly. I've attached a plot below. The channels are H1:TCS-ITM{X,Y}_HWS_DALSACAMERASWITCH.