Sheila, Oli, Jennie, Tony, Ryan Crouch
Summary: Today we found that the 9MHz modulation depth was low because of a wrong setting, and that the beam was misaligned on the LSC REFLAIR A diode. After these changes and some gain adjustments we were able to lock DRMI.
Details:
After restoring the 9MHz modulation depth this morning (89497), we went on to locking PRMI, using the notes from 89491, but scaling for the increased modulation depth. Once it was locked, Oli and I went to ISCT1 to double check the alignment onto the POPAIR diodes, we slightly improved the alignment onto POPAIR A, but POPAIR B was already well aligned. We measured 0.55mW in the POPair path before the splitter using a power meter.
Oli phased POP18 + POP90 to maximize the I ERR signal with PRMI locked, and we put a 132 Hz, 3000 count exitation into PRM LOCK L to phase REFLAIR 9 and 45 to minimize the PRM signal in Q. (userapps/lsc/h1/templates/phase_REFLAIR.xml) We then adjusted the gain in REFLAIR 9 to match the ugf to the PRMI reference (screenshot). In O4, we used a whitening gain of 12dB and then a digital gain of 0.25 before the phase rotation, and LSC input matrix of 1.75 . Now we are using 30dB of whitening gain and a digitial gain of 4.4 to reproduce the same ugf (input matrix unchanged 1.75), so we are missing a factor of 140 signal strength compared to O4. For MICH the ugf is also right on the reference, screenshot. This is with 30dB of whitening gain and a digital gain of 4 before the phase rotation, compared with 0dB of whitening gain an a digital gain of 1 in O4, so this signal is too small by a factor of about 126. So this is pretty similar to the signal loss for REFLAIR 9.
Oli and I went back to ISCT1, removed the ND filter, and realigned reflair A. This gave us 2.5 counts on REFLAIR A LF, and we were able to relock PRMI with the whitening gains for REFLAIR back to normal. We adjusted the phase of reflair, and added some digital gains to set the ugfs to what they used to be.
A comparison time with PRMI locked in November 14th 2025, 22:23:00 UTC
| Nov 14th 2025 22:23 UTC PRMI locked | today PRMI locked | nov/now ratio taking gain changes into account | |
| POPAIR A LF | ~17 counts | 16 counts | fine |
| POPAIR B LF | 12 counts | 8 counts | 1.5 |
| POPAIR B 18 | 100 counts, whitening gain 12dB, digital gain of 2 | 80 counts, whitening gain is 36dB, digital gain of 2 | 20 (improved with aligment) |
| POPAIR B 90 | 70 counts, whitening gain 6dB, digital gain 4 | 90 counts, 18dB whitening gain, digital gain 4 | 3 |
| ASC POP X DC SUM | 40 counts | -1 | running raster on PM1 |
| reflair A LF | 2.5 | 2.5 | after today's realignment |
| reflair B | 5.5 | 5.3 | yay? |
| reflair A RF9 signal | whiten gain of 12dB, digital gain of 0.25 | whiten gain 12dB, digital gain of 0.31 | 1.25 |
| reflair RF45 signal | whiten gain 0dB, digital gain 1 | whiten gain 0dB, digital gain of 2.17 | 2.17 |
We ran SRY alignment manually, (the WFS don't work since we have not rephased them).
This screenshot shows the gains I had manually adjusted when we got a 3 second DRMI lock, March 13th at 23:28:21 UTC.
DRMI locking references: trigger levels: 44348 filter states and loop measurements: 87768 When DRMI grabs for a few seconds, it looses lock due to a 20Hz oscillation (screenshot), which suggests the MICH gain is too high. I tried a few changes to MICH gains. In this grabbing attempt, it seems that the MICH filter changes are what is making things unstable.
Edited to add: We have locked DRMI, with three manual changes, to the POP18 digital gain (set to 3 as in screenshot), RF9 gain set to 0.31, and RF45 gain set to 2.17. We think that this was too high for RF45, and instead it should be left as SDF will set it at 1. We stepped through the DRMI_locked guardian state manually which was fine.
