Sheila, Jenne, Keita, Georgia, Danny, Rich, Jeff and Evan

ALS working in rainy weather:

It's raining and we have locked the interferometer several times.  We need more statistics before we make a conclusion about the ALS glitches being fixed after the fiber was burried, but we are grateful that there is more dirt on it. 

ASC:

To get back to stability as quickly as we could, we reverted the ASC changes described in 46729, and made a few other changes.

• we brought back some of the bost10W filters which were really a series of 4 resGs in DHARD P, DHARD Y, and CHARD P.
• We also added back in the 0.47Hz resG in DSOFT P.
• Keita and I made a resG like boost for DHARD P (1st screenshot), which should add gain at 1 Hz where we have an unexplained dip in the transfer function, and shift the 0 phase margin point up slightly.  This seems to work.  The second attached screenshot shows the DHARD P OLG with the new boost on, before we reduce the gain from 50 to 30 and engaged the cut off filter.  (yes, that is a total of 14 resG filters) Sometimes a point design works for a weekend
• We also put the RPC for the soft loops back into the gaurdian.
• Even with these changes we still had problems with 0.43-0.44 Hz instabilities in DSOFT P, and CSOFT P, including one lockloss.  We increased the gain of both DSOFT and CSOFT P by a factor of 2, that is in the guardian.  We tried to disengage the 0.47 Hz resG one time when this was ringing up, but this caused a lockloss.
• We tried reverting the SRC2P in the guardian, so that it was engaged with a gain of 100 in full lock and then reduced before power up.  However, this caused it to oscillate so we have left it engaged with a gain of 15, which is still used in low noise. 
• We had some trouble with the DRMI ASC, so we have taken it out of the guardian. We have sometimes been aligning by hand in DRMI before going through the lock acquisition sequence. 

Changing the DARM actuator:

• We lost lock transitioning the DARM control from IX ESD plus EY L1L2 to ETMX L3L2L1 3 times tonight.  The first time it was done by the guardian in the usual way (we weren't expecting problems here since it has been fine for several weeks). 
• The second time we tried our usual trick of measuring the DARM loop, transitioning half way to a new actuator, and remeasuring , but even the half transition rung up a 3 Hz instability too quickly for us to recover.
• In the next lock, we made comparison measurements of the two ESDs response in DARM by exciting in the DRIVEALIGN matrix here.  We reduced the ETMX drivealign gain from -40 to -35.7 to better match the ITMX response.  Georgia put this in the guardian. 
• We also made comparison measurements between the combined L1+L2 actuator, by exciting the L2 LOCK L filter when the suspension is set up to make the transition.  We measured this from about 30-10 Hz, around the ESD crossover with the PUM, it matched well for the two arms with a 180 sign flip.
• We were able to transition the ESD control from ITMX to ETMX with the reduced drivealign gain.  We attempted to transition L2 from ETMY to ETMX next, we transitioned 50% of the way, the loop shape seemed fine around 20Hz, but the interferometer had large glitches.  We lost lock very quickly with a 3Hz instability when we tried to transition to 75% ETMX and 25% ETMY.
• Georgia edited the guardian so that it will leave L1+L2 control on ETMY and transition the ESD control to ETMX.  We did these steps by hand in the next lock by copying and pasting from the guardian, so this should be fine for the next lock.  
• We do not know what caused this problem, we have checked conlog, SDF, and the guardian SVN and nothing in the settings seems to have changed.

Other notes about this locking configuration:

• With this change to the DARM actuator, our calibration lines are not well matched to the Pcal lines, especially at 35Hz.  We are not planning to do a pcal sweep this weekend, but one can be done before we change the configuration.
• We originally changed the DARM actuator because of large scattering shelves that we saw while actuating on ETMY.  So far tonight we haven't seen the scattering problem and the glitch rate seems comparable to other recent locks. 
• We have high coherence between DHARD and DARM up to 50 Hz, this is because of our ongoing trouble with decoupling the angular noise after moving the dither lines below the DARM band.  For tonight we aren't going to try to address this coherent noise which could be subtracted if needed.

Tips for re-locking:

• If we loose lock we might need to do an initial alignment before re-locking, this afternoon and tonight we have done a lot of manual alignment tweaking which is working OK.
• We had one lockloss tonight due to the OMC guardian setting the H1:OMC-READOUT_ERR_GAIN to a value smaller than normal.  It might be a good idea to check that this gain is around -2e-6 before transitioning to DC readout until we have a chance to understand why this happened.
• We have been stopping in check DRMI ASC and aligning manually before doing the CARM offset reduction. 
• Engaging the ASC has been rough, so before letting it engage we have been sitting in PREP_ASC, and watching the build ups and error signals while we use hang's arm move script to bring the CHARD error signals close to zero before closing the loops.
  • Hang's script is /ligo/home/hang.yu/Desktop/ASC/ARM/move_ARM_dev.py CH P 0.03  will step CHARD P in the positive direction by 0.03 urad (ish)
• When the ASC comes on the POP18 build up is dropping and the POP 90 is increasing which can be alarming, things recover somewhat when the dither loops come on.