We have achieved resonance, and could probably make it to carm analog. Unfortunately, the steps are not fully automated just yet. Here is a summary of how we are doing this:
- Lock to "engage DRMI ASC"
- the SRC Y ASC works, but the SRC P ASC does not yet work, so I take us here, manually engage SRC Y, and then adjust SRM pitch by hand
- MICH ASC is disengaged after DRMI ASC offloaded
- Louis and I attempted to get MICH ASC running on AS 36, we failed twice and gave up in favor of getting to resonance
- The beamsplitter oplev servo should be re-engaged at this point, pitch gain is 300, yaw gain is 650. It can be done by hand
- We had some locklosses due to oscillations which turned out to be the beamsplitter moving too much, the oplev servo helps this
- From here, we can run through to CARM OFFSET REDUCTION
- At this point, the current arm alignment is very far off from the zero point of the DHARD error signals.
- I was using the move_arms_dev.py script to walk the DHARD error signals to zero. Steps of +- 0.01 work well. Turn off the smooth limiter for both
- Then, DHARD WFS can be run
- Earlier today, after we let DHARD converge, I reset the green camera references. This means that if we run a full initial alignment, this step should no longer be necessary. I recommend doing this asap. Tagging opsinfo
- We have had some trouble running carm 5 pm with the guardian
- Both Sheila and I have been running through this state by hand. I think I have edited the guardian to work properly, but if it doesn't, I recommend the following steps:
- Run lines 2747 and 2752 in ther ISC_LOCK guardian to raise the PRCL gain
- Adjust the TR CARM offset by hand up from -40 to -48. As you do, remeasure the tr carm OLG and check that we are between the two blue reference traces. If the gain is too low, increase the TR CARM gain. I have found that 1.5 works well, this is what is in the guardian
- We can then run carm to refl and resonance in the guardian
- Louis and I think we have adjusted the analog gain settings in carm to analog such that it should run without intervention, but we haven't tested it yet.
I know Sheila had her own alog documenting our work for most of the day, but I don't see her post just yet.
I'm attaching the green camera SDF, and also the rephasing of RF36 that Louis and I did earlier.
notes on continuing to get back through to CARM_TO_ANALOG:
- first try: lost lock at CARM_TO_TR
- I asked Ibrahim to run initial alignment for green
- asked Ibrahim to actually go ahead w/ full initial alignment while I moved to home office
- second try (first after running initial alignment): lost lock shortly after getting to CARM_OFFSET_REDUCTION before I was able to take a look at whether I should walk DHARD.
- third try: lost in CARM_TO_TR
- fourth try: went to START_TR_CARM and lost it before getting a chance to take TR_CARM OLG measurement
- fifth: went to prmi_asc then back to ENGAGE_DRMI_ASC
- Sheila joins online
- also added oplev damping to guardian in ISC_LOCK.py:OFFLOAD_DRMI_ASC
- paused in DRMI 1F to adjust alignment, checked MICH OLG, OLG looked bad ... in the way that it often does when there are mode hopping problems. I didn't get a screenshot unfortunately.
- mode hopping problems are back. alignment is not great. we reverted the alignment back 4PM PDT using the restore button on IFO ALIGN Compact screen. This reverted the alignment for all the optics on the COMPACT IFO ALIGN medm screen I think.
- I reset the green camera references to what they were before Elenna updated them this afternoon. We're not sure what is wrong but suspect that the alignment got worse in the evening.
Morning operator: We are leaving the IFO in DOWN. please run a full initial alignment when you get in.
Sheila, Elenna, Louis, Oli
This morning Elenna and Louis looked using AS36 for BS ASC.
I added a guardian state called SERVO_COMM offset, which looks at the TR_CARM inmon and servos the ALS COMM VCO tune offset to bring it close to 0.5. I'm hoping this will help make the TR_CARM transition more reliable by making sure we start closer to the correct offset, it has worked (5 out of 6 times), we will see.
We had several more locking attempts, stepping up the CARM offset reduction.
Elenna engaged DHARD WFS by hand partway through the CARM offset reduction state, TR offset -12 I think. We stepped up the REFFL BIAS gain in steps at TR_CARM of 40, rather than waiting for 52 as the guardian would have. We lost lock going to TR_CARM of 50, although it looked like we had enough gain margin in TR_CARM.
Next lock: Let guardian take us to DHARD WFS automatically. PRCL ugf was about 30 Hz, we changed the input matrix to 4.12 to get the ugf to 52 Hz. stepped TR_CARM offset to -47, increased TR_CARM gain to 1.3 (guardian would have done -52 offset and 2.1 gain).
Next lock: Stepped TR_CARM offset to -48, increasing gain to 1.5 in steps as we increased. We could then see that we were in range of the TR_REFL error signal, and measured the transfer function using the TR_CARM template. We were able then to run through the CARM_TO_REFL and RESONANCE guardian states as written, but lost lock after 1 minute at resonance.
Later this evening we've had difficulty, seemingly related to being in a different alignment (both mode hopping issues and trouble with TR_CARM).