Virgo Logbook

AdV-ISC (Commissioning up to first full interferometer lock)

casanueva, nardecchia - 16:46 Sunday 20 August 2023 (61316)

Lock acquisition recovery + SRCL SET loop study

The plan for today's shift was to try to re-commission the SRCL SET loop, since it might have been misfunctioning for a while (#61313, #61308). However the first trials to lock were all systematically unlocking at 400mW due to a differential oscillation (probably DARM) at 1Hz or 75Hz, even if the gain was only changing from 0.015 to 0.025. So After trying adjusting the gains of DARM I contacted Ilaria (while I also tried to test if the situation improved by closing the CH flips at different steps of the lock acquisition). Se checked the status of the CO2 lasers and adjusted one of the rings by 15%. We also agreed that the best spot for turning off the CHs is after a ramp, so I put it at STEP 2_3.

After this action we reached CARM NULL 3f without major problems, and I disabled the SRCL SET loop in the automation. In a first lock I adjusted maNually the SRCL SET and injected DARM noise in order to cross-check the goodness of the present error signal (starting 10.56.04 UTC). Then I tried to reach LN but failed on the transition from LN1 to LN2.

After I added the possibility of injecting a 4th line in DARM, called DARM_LLF_line. I also decreased the UGF of the SRCL SET loop and I engaged it manually, while turning on the line at 30.3Hz (15.37 LT). This way we can see if putting a line at a different frequency region would give us a better quality error signal for the OS.

During the locks the sidebands were quite balanced (or at least not diverging as last week), but there were mini-jumps happening often. Maybe causing unlocks from time to time, the reason for the unlocks need to be better understood. Figure 1 shows the trend of the main powers after Ilaria adjusted the DAS.

Since the lock acquisition has been very unstable during the past days, I let the SRCL SET loop engaging at the last possible moment of CARM NULL 1f and with a lower UGF. And I let the ITF to relock in CARM NULL 1f only (I think).

Images attached to this report

Comments to this report:

casanueva - 9:15 Monday 21 August 2023 (61320)

Yesterday I left the ITF locked in CARM NULL 1f, and it remained locked about 7 hours. At the end, there was a jump from the state with the sidebands unbalanced to the one with sidebands balanced, and not long after the ITF unlocked.

After this unlock, the ITF tried to relock but failed for more than half an hour to lock the DRMI and then it went to LOCKED ARMS.

Images attached to this comment

ruggi - 11:08 Monday 21 August 2023 (61321)

The long lock had a trouble after 6 hours: the vertical mode of PR mirror (the old 11.5 Hz, fig 1) started to grow, spoiling about all the ITF signals. In particular, the corrections on PR, BS and SR MIR became very high, close to the saturation. This fact could have induced some ring heater - like effect, changing the optical properties of the central cavities. The other easier guess is that the bad status of the signals induced offsets on the alignment, but I think we should not discard the thermal effect as a trigger of the jump.

After the jump, the coupling of PR vertical mode was dramatically reduced and quite soon it was back in a damped state. I have no explanation for that.

Images attached to this comment

calloni, de rosa - 16:44 Monday 21 August 2023 (61326)

We gave a look to some environmental and CO2 laser signals, that we report in figure 1. The temperature of the (air of) the TCS benches is smoothly decreasing; the temperatures of the lasers are fluctuating and the laser powers are fluctuating coherently, by about 1% at highest peak after few hours. The power at the ISSN point is fluctuating of the order of 3%. We checked if these fluctuations were similar to temperature fluctuations of the benches (measured with a thermometer sensing the air temperature) or the chiller temperature but there in no correlation.

Images attached to this comment

rocchi - 0:02 Tuesday 22 August 2023 (61333)

The benches temeprature decreasing trend is due to the closing of the CH flipper mirrors.

The WI CO2 laser temperature is stable, and so is the output power.

The NI CO2 laser temperature is unstable, very likely due to the chiller not performing at best anymore (this is not evident by the chiller water temperature, which is in-loop in the cooling system), since it has been heavily stressed in June/July (60774). Like for the WI one (60809), it will have to be replaced at first occasion (when the serviced chiller comes back from the company). The power fluctuations are directly correlated to temperature of the laser head.

The NI photodiode is adding some more information: the fluctuations are anti-correlated with the temperature of the laser head and the laser is more stable and less noisy when the temperature is higher. This is an indication of the laser mode-hopping between two different states. Increasing the chiller set-point by few (2-3) cents of a degree should cure the problem.