Happened during some commissiong, unsure of the cause.

Sheila D, Chris W

While we were out of observing for other reasons, I turned off the pum noisemon filters which were calibrating the signals into DAC counts.  This was because Chris found that with the signals calibrated this way, we would need whitening to avoid digitization noise in the DQ channels.  Instead of doing that we are just removing the calibration filters.  

Accepted in SDF.  

Sheila, Jenne, Rahul

We are implementing new Guardian settings in the python code for damping the violin modes. Till now Guardian use to switch off damping for any of the modes rising by 1 order of magnitude. In the new settings, Guardian will just flag the following message (give below) and will continue damping them.

'mode %s is growing!! not turning off gain'

We have found that during Monday CAL sweeps (modes rise slowly over few hours) or other days whenever the modes were rung up (often the monitor levels were below 1, and they are just bouncing up and down), Guardian use to switch them off. We feel this is not necessary as those modes with their correct damping settings should be successfully damped.

I will keep on eye on the new settings (especially on Monday when Jeff. K will perform CAL sweeps) and will report the progress.

The updated VIOLIN_DAMPING.py has been committed to the svn.

TITLE: 03/13 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Observing at 117Mpc
OUTGOING OPERATOR: Corey
CURRENT ENVIRONMENT:
    SEI_CONF state: WINDY
    Wind: 5mph Gusts, 3mph 5min avg
    Primary useism: 0.01 μm/s
    Secondary useism: 0.17 μm/s
QUICK SUMMARY: Locked for almost 2 hours after some trouble and a lock loss. Calm environment.

Summary:  At about 8:30utc, H1 took a 20Mpc drop in range & hovered there for ~4hrs until it dropped out of lock.  Cursory look didn't show anything obvious.  H1 was able to get back to NLN without intervention in under an hour and the range appears to be normal-ish so far.

Timeline:

• 8:37utc (1:37amPT) H1 range begins to drop over a few minutes from ~119Mpc down to 102Mpc (no notification received, other than Richard text at 13:03), and stays here until...
• 12:24 LOCKLOSS
• 13:19 NLN
• 13:22 OBSERVING...range is hovering around 116.5Mpc (after first few points around 119Mpc)

Quick Check:

Did a quick check on a few items to see about reasons why H1 took the drop to 102Mpc, and looked at "usual suspects":

• DIAG_MAIN:  NOTHING observed in the log.
• SQZ_MGR:  NOTHING observed in the log.
• ISC_LOCK:  NOTHING observed in the log.
• SUS Violins:  NOTHING observed rung up above 3 for fundamental and 1st harmonic.

TITLE: 03/13 Eve Shift: 23:00-07:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Observing at 119Mpc
INCOMING OPERATOR: Corey
SHIFT SUMMARY: Nothing to report
LOG:
 

TITLE: 03/12 Eve Shift: 23:00-07:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Observing at 114Mpc
OUTGOING OPERATOR: TJ
CURRENT ENVIRONMENT:
    SEI_CONF state: WINDY
    Wind: 11mph Gusts, 7mph 5min avg
    Primary useism: 0.02 μm/s
    Secondary useism: 0.19 μm/s 
QUICK SUMMARY:

TITLE: 03/12 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Observing at 115Mpc
INCOMING OPERATOR: Jim
SHIFT SUMMARY: We had a few hours of commissioning, but stayed locked. Current count is 23 hours.
LOG:

• 1724 Out of Obeserving for commissioning
• 1742 Jeff K and Jeff visitor to CER
• 1800 Jeffs out
• 1826 Robert to EY to get ready for his commissioning work
• 1930 Observing

Ive added a state in IFO_NOTIFY for when we receive a SNEWS alert. This will work in conjunction with Dave's LLA system to make the proper notifications when necessary. Upon receiving a SNEWS alert on site, this node will stay in that state for 30min. At this point it is up the the operator to follow the normal Site Response to Astronomical Alarms (L1500117) procedure.

[SudarshanK, SheilaD]

We were able to stay locked with lower DHARD Gain last time (LHO alog #55526) ans Sheila made some changes to the DHARD filter yesterday ( LHO alog #55556). We took a regular sensing function measurement  to make sure the filter change has not affected our calibartion. We also made another sensing function measurment with a lowered DHARD Pitch gain (from -30 to -10). Both  measurements are saved at the following location below. Analysis to follow.

aligocalibration/trunk/Runs/O3/H1/Measurements/FullIFOSensingTFs/2020-03-12_H1_DARM_OLGTF_LF_SS_5to1100Hz_15min.xml

aligocalibration/trunk/Runs/O3/H1/Measurements/FullIFOSensingTFs/2020-03-12_H1_PCALY2DARMTF_LF_SS_5t1100Hz_10min.xml

aligocalibration/trunk/Runs/O3/H1/Measurements/FullIFOSensingTFs/2020-03-12_H1_DARM_OLGTF_LF_SS_5to1100Hz_DHARDGain10_15min.xml

aligocalibration/trunk/Runs/O3/H1/Measurements/FullIFOSensingTFs/2020-03-12_H1_PCALY2DARMTF_LF_SS_5t1100Hz_DHARDGain10_10min.xml

On Tuesday I measured the dark noise of the OMC DCPDs in various analog electronics configurations, to help understand what our options might be for running in a different configuration that would allow us to use a higher DARM offset. 

In the attached screenshot I have the dark noise of 4 different configurations, high and low transimpedance and high and low analog whitening gain.  Although unlabeled, the yaxis is in mA/rtHz.  These are at OMC DCPD SUM, so have digital compensation for the analog changes: the High Z filter module is off for the Low Z measurements, and the digital gain is reduced by 24dB for the measurements with increased analog gain.  The increase in analog gain has an impact in the region where we are limited by the ADC noise, below 12 Hz or so.  Above about 12 Hz, the dark noise increases by a factor of 2.5 when we reduce the transimpedance by a factor of 4. 

The noise limit at higher frequencies for the LowZ case is well explained by noise from the DCPD itself, which Rich Abbott has quoted to me as 300 nV/rtHz at 100 Hz over the low Z 100 Ohm resistor coming into the whitening chassis.  This translates to 3.02e-8 mA / rtHz and matches my spectrum at 100 Hz. If I assume that the 300nV/rtHz over the 100 Ohm resistor is a current noise, then that gives me 3 nA/rtHz.  Assuming that that current noise is constant then over the higher impedance 400 Ohm resistor I should have 1200 nV/rtHz coming into the whitening chassis.  But, that would result in 2.15e-8 mA/rtHz at 100 Hz, which is higher than what I measure.  So, I think I need to check with Rich to see if the DCPD noise is different somehow over the higher impedance.

Anyhow, Rich and PeterF have looked at some noise simulations, and it seems like we can reduce the analog whitening gain by a factor of 2 and leave the DCPDs with the higher transimipedance, so we'll likely pursue that avenue.

In the attached .xml file, I have the several references saved, in case they are ever helpful to come back to.  For each case, I have the OMC DCPD sum, as well as the IN1 channel of each PD.

1. Refs 0, 1, 2: Nominal configuration. 0dB analog gain & digital gain = 1000, analog HighZ & HighZ filter module on, 2 whitening filters and 1 lowpass in analog & compensating digital filters.
2. Refs 3, 4, 5: Same as set #1, but increased analog gain. +24dB analog gain & digital gain = 63.096 (-24dB from 1000), analog HighZ & HighZ filter module on, 2 whitening filters and 1 lowpass in analog & compensating digital filters.
3. Refs 6, 7, 8: Same as set #1, but removed 1 whitening filter. 0dB analog gain & digital gain = 1000, analog HighZ & HighZ filter module on, 1 whitening filter and 1 lowpass in analog & compensating digital filters.
4. Refs 9, 10, 11: Same as case #3, but also removed lowpass.  0dB analog gain & digital gain = 1000, analog HighZ & HighZ filter module on, 1 whitening filter in analog & compensating digital filter.
5. Refs 12, 13, 14: Same as case #4, but increased analog gain. +24dB analog gain & digital gain = 63.096, analog HighZ & HighZ filter module on, 1 whitening filter in analog & compensating digital filter.
6. Refs 15, 16, 17: Same as case #2, but with LowZ.  +24dB analog gain & digital gain = 63.096 (-24dB from 1000), analog LowZ & HighZ filter module off, 2 whitening filters and 1 lowpass in analog & compensating digital filters.
7. Refs 18, 19, 20: Same as case #6, but with lowered analog gain. 0dB analog gain & digital gain = 1000, analog LowZ & HighZ filter module off, 2 whitening filters and 1 lowpass in analog & compensating digital filters.

Out of Observing at 1724 UTC. Planned to be out for around 2 hours.

There was a small earthquake this morning that passed through, which I expected to have cause the seismic system to transition to the earthquake state. It didn't, and probably didn't need to, but it did indicate that the seismon test in SEI_ENV was too strict. That test stopped checking whenever the seismon arrival time went to 0, so wouldn't trigger on an earthquake if arrived after the seismon prediction. There's an independent "passive" test that triggers when peak mon goes above 1000 nm/s, but for moderate earthquakes, that could be too late to prevent a lockloss.

I've changed the test to give us a window of 2 minutes after the seismon prediction. So line 49 used to be:

           ezca['SEI-SEISMON_LHO_R35_ARRIVALTIME_TOTAL_SECS_{}'.format(i)] > 0:

is now: 

           ezca['SEI-SEISMON_LHO_R35_ARRIVALTIME_TOTAL_SECS_{}'.format(i)] > -120: