TITLE: 08/19 Day Shift: 1430-2000 UTC (0730-1300 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: Oli
SHIFT SUMMARY: GV6 & 8 were opened this morning, so we now have a full IFO to commission. Commissioners have been hard at work aligning to the arms. 
LOG:                                                                                                                                                                                         

Louis and I immediately started where we left off yesterday with BS M3 commissioning. Louis copied over the LLO filter design that sets up the BS feedback with a 7 Hz crossover between M2 and M3, 82200 (thanks LLO for doing the heavy lifting here!).

I copied these filters into my matlab model so we could confirm we understand what it is doing. This plot shows the modeled feedback with the locking and drivealign filters applied. The LLO design is applied to M2 and M3, such that the crossover is roughly 7 Hz as expected. I applied our current M1 offloading scheme, and the M1/M2 crossover is still 30 mHz.

This looks good to us, so we proceeded to copy over the filters to the BS sus banks. I first locked MICH dark using the old feedback design without the lowpass filters so we had much more phase. This provides a reference for us to make sure we do the switchover correctly.

Louis and I engaged the correct filters in ISCINF M3 L, M3 lock L, M3 drivealign L2L and M2 lock L. Here is a screenshot of the old M2 filter design, and here is a screenshot of the settings we used to test the new feedback. I want to note that one difference is I found that we need a gain of -1 overall in BS M3 drivealign to get the phase correct.

I did touch up the MICH alignment before locking, since the arm team is working on aligning the arms now that gate valves are open (spoiler alert).

We successfully locked MICH dark, no issues! I measured the MICH olg again, and saw that we were almost exactly where we wanted to be in the gain, and the phase is a little different but not drastically so.

Masayuki noted in a previous alog that the MICH gain is a bit low, it would be better to have a 10 Hz UGF. Therefore, I bumped up the gain by a factor of 3 and remeasured. Looks good! This plot compares all three MICH dark olgs.

Louis has updated the ALIGN_IFO guardian with these new settings and is working through ISC_DRMI. In addition to the filters in the BS sus, we are updated the MICH dark and MICH bright gains to be 3x higher (so plus/minus 7200 instead of 2400).

As a note, ALIGN IFO does not use M1 feedback, we expect to test this when we lock PRMI/DRMI. However, I don't predict any issues, as the crossover should still be the same.

I have accepted new settings in the SDF as well, screenshot.

Below are the foton strings that we copied over:

BS_M2_LOCK_L

FM1: zpk([-4.398229715026552+i*5.654866777052986;-4.398229715026552-i*5.654866777052986], [-344.8589087698574+i*562.4185982606715;-344.8589087698574-i*562.4185982606715; -2513.274122871833],21314144.89497595)

FM2: zpk([0+i*754.0107241782374;0-i*754.0107241782374], [-114.2660735869546+i*205.9572881820654;-114.2660735869546-i*205.9572881820654], 0.09757621338967941)

FM4: zpk([],[-50.26548245743591],50.2654824574367)

As promised, this morning I took a new measurement of the JAC reflected beam on wavefront sensors A and B. This time, we unlocked the JAC, which provided a cleaner beam for centering in both pitch and yaw. The results were much better, with the relative differences between segments reduced to below about 10%.

As a sanity check, we compared the measured and calculated G1/G4 ratios for WFS A and WFS B, as shown in the attached table. For WFS A, the discrepancy is about 10%, while for WFS B, it is only about 2%. We think the remaining error may come from beam clipping when we attempted to recenter the beam between the two halves, the approximately 1 mm gap between the quadrants, or pitch-yaw cross-coupling.

Overall, this measurement gives us an estimate of the uncertainty associated with this technique for measuring the relative segment gains.

Long day, short alog

We continued to try to lock PRMI today. After consulting with Sheila, we moved IMs1-3 back to the alignment they had during the arm peek last week. Ryan S has noted the osem values for each alignment, so we can switch back and forth between these two configurations. Then, Louis and I reverted IM4, PR2, PRM, SRM, SR2, BS to their alignments during a DRMI lock last week. I also moved SR3 to its current alignment (TCS team had moved it back to old O4 alignment for HWS work).

This alignment was very bad for flashes, so I iterated between initial alignment work on PRX and MICH, and running a PRMI lock. We found that it was very difficult to lock PRMI, both because the alignment was poor and the beamsplitter was moving a lot. Once we turned back on BS oplev damping, PRMI locking was more frequent due to less BS motion, allowing us to better touch up the alignment. However, there was then a large 6 Hz oscillation. After checking many things, it turns out the oplev damping is causing this oscillation, both PRCL and MICH are stable. So, we need oplev damping to lock but we can't keep it on or we will unlock.

After talking with Sheila again, I moved to try commissioning the BS M3 feedback as a way to help this problem. Louis and I tried measuring the beamsplitter plant with the offloaded feedback to M2 and M1 and then M3 and M1. We found that we needed about a gain of -300 on M3 to replicate the M2 plant. However, we could not lock MICH dark with this feedback.

We tried many different things after this which I won't detail, but Keita and I performed a different type of measurement that confirmed the work Louis and I did earlier was correct. When we locked again, we saw that the issue is that the DAC rails. Looking at the master outs on M3, the saturation occurs below 10 Hz. I see that we should replicate the LLO feedback that is distributed between M2 and M3 with a crossover of 7 Hz. See 82200. I will pick this up tomorrow.

Editing to add spectrum of BS M3 master out with length feedback applied and the appropriate gains. I put a horizontal line roughly at 134e6 DAC counts, which is where we expect the saturation to occur on our 28 bit DACs. A crossover of 7 Hz as LLO has would work well. Louis is copying their filters over now.

TITLE: 08/19 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 4mph Gusts, 1mph 3min avg
    Primary useism: 0.01 μm/s
    Secondary useism: 0.07 μm/s 
QUICK SUMMARY: Dust alarms in the PSL overnight, but they quickly dissapated. There was a verbal alarm at 0828UTC "Check corner station HEPI pumps", I don't initially see any issues but I'll continue to investigate.

Plan today is continued corner station commissioning unless we get the arms opened.

TITLE: 08/19 Eve Shift: 2330-0500 UTC (1630-2200 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY:

IFO is in IDLE with IMC Locked and Power at 2W

Evening was spent trying to comisson BBSS M3 stage as a way to aid with alignment. Keita, Louis and Elenna have some next steps to try tomorow. See alog 91593.

After they were finished and per Jennie's instructions I changed the JAC Heater POWER_SET from 1.465 to 3. Then, I informed Huyen that Seismic overnight tests can begin.

LOG:

Since enabling irqbalance on dc0 we seen less CRC errors (only around restart times for the last few days).

Today we enabled irqbalance on dc1.  Unlick on dc0 there was a small (but visible) dip in the average cpu use.

We will continue to watch this.

TITLE: 08/18 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: Ibrahim
SHIFT SUMMARY: Were were able to get PRMI locked again today for a bit after more alignment work, and then some commissioning of the BBSS was done after that. The HWS and CO2 systems were worked on and some 3IFO checks were done as well. Dust alarms have been going off most of today in the optics lab. The LVEA remains Laser HAZARD.
LOG:

Jennie Wright, Masayuki Nakano, Khanh Vu

Masayuki’s Math

We previously made an attempt to balance the JAC’s reflected beam on the WFS quadrants. However, Masayuki realized that this might not be a fair measurement because the voltage is defined as Voltage = Gain x Power, and the gain can be different for each segment.

To obtain a quantity that can be compared between segments, we instead excite the beam. For example, we can move the beam to the upper half of the WFS (Segments 1 and 2) so that no light lands on the lower half (Segments 3 and 4). If we then drive the beam in yaw with a sine wave, the total power on Segments 1 and 2 must be conserved. Therefore, any change in power on one segment results in a corresponding negative change on the other segment.

The powers on Segments 1 and 2 can then be written as:

P1 = P1_DC + ΔP sin(wt)

P2 = P2_DC - ΔP sin(wt)

Since Voltage = Gain x Power,

V1 = G1 x (P1_DC + ΔP sin(wt))

V2 = G2 x (P2_DC - ΔP sin(wt))

At the excitation frequency, the amplitudes of the power changes on the two segments are equal and opposite. Therefore, taking the ratio of the measured AC amplitudes gives the gain ratio G1/G2.

Applying the same procedure, we can measure G1/G2, G2/G3, G3/G4, and G1/G4. However, we can also infer G1/G4 from the first three ratios:

G1/G4 = (G1/G2) x (G2/G3) x (G3/G4)

Comparing this inferred value with the directly measured G1/G4 gives us a consistency check.

Measurement Procedure

We performed this measurement by injecting an excitation into the JM1 suspension system. Before doing so, we opened the ASC loops and paused the JAC Guardian. We used the picomotors to move the beam onto different halves of the wavefront sensors.

For precision, we monitored the DC signals from the individual segments. Using the same example, when positioning the beam on the upper half, we moved it until there was essentially no signal on Segments 3 and 4. We then balanced the beam in yaw so that the DC signals on Segments 1 and 2 were approximately equal. An example dtt for this step is attached below.

The excitation was a sine wave at 11 Hz with an amplitude of 12. For the upper and lower halves, we injected in yaw. For the left and right halves, we used pitch excitation. The power spectrum plot is also attached below.

Results

We measured the segment responses at 11 Hz and divided the corresponding amplitudes to obtain the gain ratios. The results are shown below.

For WFS A, the directly measured G1/G4 does not agree well with the value inferred from the other three ratios. For WFS B, the directly measured and inferred G1/G4 values agree considerably better.

However, the calculated results themselves do not seem physically reasonable, since they imply that the gains of different segments can differ by as much as a factor of 10.

Next Step

We suspect that this issue comes from performing the measurement while the JAC was locked. In this condition, the reflected beam contains significant higher-order-mode content, making it difficult to position and balance the beam cleanly on the WFS quadrants.

We will therefore repeat the measurement tomorrow with the JAC unlocked.

JAC Heater ran out of range this afternoon and the JAC unlocked. We started this lock stretch (first x cursor) already at a pretty high PZT offset so I might add a upper limit to how high a voltage we should lock at.

I have increased the gain magnitude in the heater control servo from -0.1 to -0.15 to try and stop this happening again before this evening when we will increase the power set value from 1.5 W to 3W to give us more headroom above ambient temperature for the servo.

TJ, Fil, Camilla

We turned on CO2X laser without issues. TJ had to turn on Outputs of power supplies on the mezzanine. We searched for home of the rotation stage and checked that the beam was aligned to both irises before the periscope. Leaving laser on at minimum power. Guardian did successfully lock laser. 

For CO2Y, we were unable to turn it on. We checked SMA connections on laser head, BNC and TNC connections between RF driver and laser and tried turning on in PWM mode with the PWM driver. Fil then swapped chassis D1700745 to a spare. This did not help. TJ then swapped the RF cables t the TCS-X rack for X and Y. X laser stayed working, Y laser still not working. Fil checked the AA chassis. Fil and TJ looked at swapping the power supplies at the mezzanine but this was complicated and the power supply appeared to be behaving normally. 

TJ, Camilla

I put SR3 into it's O4 alignment (P 458, Y -135) changed from (P 524, Y -135), checked no saturating pixels (ITMX stayed at 1Hz frame rate, ITMY increased to 5Hz), took new HWS references and started the HWS code.

We are running a ring heater test with 2W/segment for 4 hours this evening. Using scripts below on tmux sessions:

• python3 /tcs/common/scripts/power_adj_scripts/ring_heater_schedule.py ITMX -s 'now' -d 4 -p 2.0  -ep 0.4
• python3 /tcs/common/scripts/power_adj_scripts/ring_heater_schedule.py ITMY -s 'now' -d 4 -p 2.0  

Ring heaters should be back to nominal by ~1am. Purpose of this is to check HWS alignment before we align the CO2s (using HWS). 

(Travis, Gerardo)
We removed the old gauge a BCG-450 and installed a new one a BPG-552.  Since the gauge at this location does not play role with relays and such, the old pigtail did just fine, provided the power to the new gauge.  The new gauge is powered on and connected to the same EtherCAT cable as the one that that we removed.  Dead volume of this assembly is getting pumped down with a small can turbo pump and an aux-cart, and a very long flex hose, and probably will continue to pump over the weekend.
No issues were encountered while replacing the gauge.