Sheila, Gabriele, TVo, Hang, Craig
Today we aligned and locked DRMI using 3f signals.
We adjusted the LSC input matrix gains to ensure MICH, PRCL, and SRCL Loops are stable. 3f gains were recorded in guardian in lscparams.py for future lock acquistion.
MICH UGF is 8 Hz, PRCL UGF is 30 Hz (potentially very little gain margin at 25 Hz), SRCL UGF is 40 Hz.
In the plots, blue is DRMI 1f locked OLGs, while red is DRMI 3f locked OLGs.
We tried closing some ASC loops, some we were only able to close with sign flips on the gain:
Pitch Yaw
=================================================
INP1 Closed with sign flip Closed with sign flip
PRC2 Occasionally Unstable Closed with sign flip
MICH Closed Closed with sign flip
Need to think harder about PRC2 P filtering, and the phasing of our WFS.
The EOM change altered our RFPD phasing. We maximized our DRMI 3f RFPD I signals by rephasing REFLAIR_B_RF27 RFPD from -80 to -85 degrees, and REFLAIR_B_RF135 from 133 to 116 degrees.
The only 3F matrix element that we modified was 135Q to MICH, which we changed from 1.25 to 2.23.
Another tricky thing is that we use an offset when we first lock DRMI to avoid mode hopping, and the guardian turns this offset off once the ASC is on and has converged. Since we skipped the ASC guardian steps today we needed to turn that offset off by hand.
Is there still a need for the offset? ITMX, SRM and the position of SR2 all have changed.
I was able to close 10 of the ASC loops, and move PRM to bring us back onto the POPA QPD.
I committed the guardian before making any changes, and have now set the guardian up to engage the ASC loops in the way that has worked tonight. This means a sign flip for both pit and yaw on INP1, MICH, PRC2, (where the pit gain is also reduced from 0.3 to 0.1) and SRC1 + SRC2 close with no sign flip.
I have changed the offsets in POP A and moved PRM pretty far to bring us onto the diode. PRM yaw seesm good with no offset on POP A, this was close to the maximum for POP18‌ build up. For pitch moving PRM down at first improved the build up but as I continued to bring the beam down on the QPD the bulid up fell. We might be able to fix this by moving PR3 while we move PRM. I have offloaded the ASC loops after doing this manual move of PRM, screenshot attached.
While everything I did is written in the guardian, just letting the guardian engage the loops didn't work I think because things are highly cross coupled and some of our loops have a lot of overshoot. Anyone who wants to close these loops can open a terminal and use guardian -i ISC_DRMI then copy and %paste a few lines at a time slowly to engage them.
In response to Daniel comment, we haven’t checked, but haven’t seen any mode hops since we got the AS camera working.