travis.sadecki@LIGO.ORG - posted 22:22, Monday 06 February 2017 (33950)
Cleared TIM error for H1IOPASC0
We are back to Observing but I am continuing to monitor the pesky PI mode 26. I worked with Terra over the phone for ~30 minutes and we were able to determine that the BP filter setting were correct for the current mode frequency (15010 Hz). I had changed them earlier from 15009 (and accepted the SDF diffs for this mode), which according to Terra should have been close enough to be effective. We were able to get the mode to slow its ringing up, but were unable to damp it down. Currently it is ringing up slowly and I expect it will break the lock again in an hour or so. I'm not sure what else I can do about this mode as all the knobs have been turned and double checked by 2 sets of eyes. At Terra's suggestion, I'll keep fiddling with gains and phases in hopes of stumbling upon the magic combination. Neither of us has seen this behavior with a PI mode before.
PI modes 26 and 9 were ringing up. Changing phase and gain weren't helping, so I started to diagnose if the BP filter had the correct values. I guess I took too long trying to figure out if the setting were correct and lock was lost.
Another lockloss at 4:33 UTC due to these two PI modes. I called Terra for some help, and she seemed to think that everything I was looking at was correct. Since this was the second lockloss due to these PI modes, I granted Terra remote access to have a second set of eyes on the issue. This time, I'm going to stop at a lower power state during locking to test the responsivity of the PI damping on the modes before ringing them up more.
Finally made it back to Observing after a bit of a prolonged IA. Had several IMC locklosses due to the PSL FSS PZT oscillating. Took quite a bit of PRM and BS tweaking to get DRMI to lock.
around 6am PST this morning the corner weather station outside temperature began working again. The plot below shows this happened when the outside temperatures went above 40F since the signal went bad about two weeks ago.
J. Kissel WP #6470 I've gathered our bi-weekly calibration suite of measurements to track the sensing function, and ensure all calibration is within reasonable uncertainty. Measurement suite was completed several minutes before the lock loss LHO aLOG 33941. The data has been committed here: /ligo/svncommon/CalSVN/aligocalibration/trunk/Runs/O2/H1/Measurements/SensingFunctionTFs/ 2017-02-06_H1DARM_OLGTF_4to1200Hz_25min.xml 2017-02-06_H1_PCAL2DARMTF_4to1200Hz_8min.xml 2017-02-06_H1_PCAL2DARMTF_BB_5to1000Hz.xml For the record, the complete suite of three measurements takes 35 minutes.
TITLE: 02/07 Eve Shift: 00:00-08:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Aligning
OUTGOING OPERATOR: Ed
CURRENT ENVIRONMENT:
Wind: 26mph Gusts, 20mph 5min avg
Primary useism: 0.15 μm/s
Secondary useism: 0.32 μm/s
QUICK SUMMARY: Lockloss at 0:28 UTC due to OMC length excitation (Sheila). I will run an IA in hopes of recoving some of the range that has been slipping away over the past ~20 hours. ISC_LOCK was reloaded (-30deg roll mode filter added to ITMX by Cheryl), in addition to the OAF model for filter changes that Sheila made.
16:05 ran a2l DTT...looks very good. It's been almost 8 hours as per TJs log.
18:06 Alarm panel in computer users room complaining about negative ground problem.
18:25 Jonathan to H2 building.
18:40 Jonathan back
21:16 Jeff into optics lab for the remainder of the day
I turned CP5 heat lamps off at lunch today. I didn't reset the heat tape circuit breaker.
Turned it back on yesterday after snow/rain fall.
I collected a roll mode power spectrum and coherence at 0.01Hz BW for all optics, and I understand the results for ETMY, ITMY, and ITMX (red, green, brown), however I do not understand the results for ETMX (blue). Roll mode peaks in ETMY, ITMY, and ITMX show corresponding coherence peaks in DARM, but the ETMX roll mode peak at 13.8828Hz is oddly low in coherence, but it has a large coherence peak at 13.9375Hz. Plot attached.
Peaks in power spectrum: ETMY, 13.8203: ETMX, 13.8828: ITMY, 13.9297: ITMX, 13.9766
Cheryl and I looked at this, the plots here are of the monitor bandpass outputs, which are a filtered version of DARM control. Cheryl has noticed that ITMX roll mode doesn't get damped to the noise floor of DARM, even with her improvement to the damping phase. We looked at the spectra of the AS WFS signals (which are used to damp these modes), and see that AS 45 A yaw is just noise, both of the AS B 45 signals have a signal for ITMX roll, but with SNRs of only 3 or 4 (the roll mode is about a factor of 20 or 30 above DARM).
We could try to damp this better by switching to either AS B or DARM as an error signal for the damping loop.
It seems that alignment has suffered a bit in the last 2 hours, or so.
22:35 Back to Observing
Typically, the amp diode powers track humidity pretty faithfully. D1 and D2, however, seem to be doing the opposite. I can't make any direct correlations for this. Everything else seems rather normal wioth with a slight downward trend in Osc chiller plots regarding flows and pressures.
Starting CP3 fill. LLCV enabled. LLCV set to manual control. LLCV set to 50% open. Fill completed in 46 seconds. LLCV set back to 14.0% open. Starting CP4 fill. LLCV enabled. LLCV set to manual control. LLCV set to 70% open. Fill completed in 1397 seconds. TC A did not register fill. LLCV set back to 38.0% open.
Raised CP4 LLCV to 39% open (from 38%).
Here is the temp plot. Striptool on the machine tracking CP 3,4 was down this morning.
Actually we found several computers and/or monitors powered down in the computer users room this morning. Logs indicate something may have happened at 10am Saturday morning.
In case this was done manually, we have put signs on the machines which support the vacuum controls group remote monitoring asking them not to be powered off.
Evan G., Joe B., Jeff K. Joe B. pointed out that there is an uncompensated ~5.215 kHz pole in the ESD output filter bank because our filter design zpk([3229.1150], [], -1, "n") results in Foton providing a pole to compensate the zero at high frequency. So the actual installed filter is zpk([3229.115],[5215.189175235227],0.9999999999999999,"n"). The ~3.229 kHz zero compensates for a pole in the analog electronics. Unfortunately, the ~5.215 kHz pole is uncompensated. Attached is a Matlab figure showing the effects from 1 Hz to 10 kHz. The band of most interest to us is <200 Hz since above 200 Hz the actuation is sufficiently rolled off that it does not contribute to the response of the interferometer. Below 200 Hz, there is very small residual magnitude error. Phase, however, is incorrect by ~1.1 degrees at 100 Hz. The equivalent delay is ~30 usec. This could explain part of the unknown actuation delay, which--for LHO--is 61 usec.
Exporting the transfer function from Foton for zpk([], [5215.189175235227], 1, "n"), we obtain a slightly different functional form than Matlab due to IIR warping near the Nyquist. Instead of 30 usec delay that approximates the phase error, a delay of 20 usec fits the Foton version more precisely. Since this effect is already absorbed into the unknown actuation delay, and any changes would require regenerating GDS and DCS filters, so we will leave this as-is. In the future, we could account for the ~3.229 kHz zero in CAL-CS model and need to be accounted for in DCS.
Associated with FRS Ticket 7351.
≈ 19:30UTC while Livingston is down.
Also: Cheryl is turning the -30˚ ITMX Roll Mode FIlter back on while we are in this state.
Sidenote: GWIstat went RED rather than Yellow? Maybe it only goes Yellow on the way UP to lock and 'thinks' we came unlocked because Intent bit went out of Observe??
Lockloss at 5:44 UTC. This one was actually caused by me putting ISC_LOCK into DOWN at Terra's suggestion. PI mode 26 was rung up enough that it was about to break lock anyhow, so Terra suggested not ringing it up any more than necessary. I'm going to wait before attempting to lock again to let the mode ring down.