I've used the attached "script" which is really just a series of commands to copy and paste into a guardian shell to transition DARM control (all stages) from ETMX to ETMY. The script is also in userapps/lsc/h1/scripts/DARM_transitions It should only be used with caution, because it relys on things being set up in a particular way before you run the script, including that the history is cleared on ETMY and that you think about the bias sign/sign in the LSC output matrix.
We were fully transitioned to ETMY at 1344452191 (Aug 13 18:56:00 UTC) with the ETMX bias at it's normal setting of -452V. I am hoping to get a few hours of quiet data in this configuration and then with the ETMX ESD bias off.
Edit: ETMX bias off at 1344457957 (20:32:19 UTC)
Edit: Switched back to ETMX bias on, darm control 4:20 pacific time
Here is a BRUCO for the normal configuration time in the above plots: https://ldas-jobs.ligo.caltech.edu/~sheila.dwyer/brucos/lho_cleaned_1344433878/
And here is a RBUCO for the time when the bias was on but DARM control was switched to ETMY: https://ldas-jobs.ligo.caltech.edu/~sheila.dwyer/brucos/lho_cleaned_DARMEY_1344452591/
There is a lot going on in the BRUCO for the nominal time
- In the nominal time, BRUCO is still finding broad and large coherence with PEM-EX_MAG_EBAY_SUSRACK_Y , this seems to be related to the coherence with ADC_0_19
- This BRUCO onfirms what we saw from the noise budget that jitter peaks are close to DARM: IMC-WFS_A_Q_YAW IMC-WFS_B_I4
- There is another group of large coherences wich seem related to the septum/ output jitter coupling: PEM-CS_ACC_HAM5_SRM_X_DQ HPI-HAM6_L4CINF_H2
- This channel seems to be similar but a bit different from the septum/jitter coupling: ISI-HAM6_CDMON_H1_V
The attached screenshot shows the DARM spectra at the normal time, with DARM swapped to EY, and with the EX bias off.
I did this test because Anamaria pointed us to 63528. When I swtiched DARM control to EY, I used some old values for the relative gains between the stages, which kept the DARM loop stable but didn't match the relative gains when DARM is controlled on EX very well. For this reason the DARM response was different, so that both the calibration and cleaning are not good in the EY configuration. There is broad and high SRCL coherence from 10-50Hz, explaining most of the high noise that we see. You can see though that the DARM peak at 75Hz and the lump at 95Hz have gone away when the EX ESD bias was turned off.
I will ask Ben and Robert for some help looking at the magnetometers at this time, but this result seems interesting enough to make it worthwhile to attempt to repeat this test while keeping the DARM loop close to the nominal so we can use the calibration and perhaps rerun the cleaning if that is needed.
Here's a screenshot that shows what's going on in this test better. The upper right shows that the noise in DARM at 75 and 95 Hz goes away when the EX bias is set to 0, but not when the DARM control was switched to EY. The two panels on the right show spectra of magnetometers at EX, they are not changed in all three locking configurations. I think this means that the peaks in the magnetometers are from something at the end station generating noise, not from the DARM control signal reaching the magnetometers. The bottom right shows the coherence between the clean channel and the magnetometers. The coherence at 75 and 95 Hz goes away when the bias is switched off, but not when DARM control is switched to EY, again suggesting that the source of this noise in the magnetometers and DARM are both from EX.
Neither magnetometer plotted here really has peaks at the calibration line frequencies, although there is coherence at these frequencies between DARM and the magnetometers. Many of the other frequencies where there is coherence between the magnetometer and DARM corespond to peaks in the magnetometer spectra, although the peak heights don't correspond well with how high the coherence is. (some of the high coherence with DARM is from peaks that are not large in the magnetometer.). This suggests that there are mulitple noise sources being picked up by the magnetometer and by DARM, with different couplings to each.
