Sheila Craig Georgia
Today locked with an even larger offset on the OMC PZT3 (PZT1). We measured the coupling from RF9 to DARM using the RF amplitude modulation injection (same as Sheila reported yesterday), and found the coupling was reduced by a factor of 2 compared to a lock we had yesterday which had smaller offsets on the PZTs (compare green and brown traces in the attachment). Craig looked at the cross-correlation between the OMC DCPDs and found the cross-correlated noise in the ~50Hz band was also lower compared to yesterdays lock.
We then changed the offset on the PZT within the same lock to try and change the RF9 to DARM coupling. We stepped the PZTs all the way back to the same configuration as two days ago (compare blue and cyan traces), and found there was little change in the RF9 to DARM coupling. So the conclusion is the RF9 to DARM coupling is changing from lock to lock, but it is not the OMC PZT offset which is causing this change.
I took some CSDs of the DCPDs to look at the correlated noise of the IFO to see whether the OMC PZT voltage changes or OMC LSC dither amplitude changes helped the coupling to DARM. During the lock tonight (Lock 3 in the legends of attachments one and two), we have seen ~15% lower DARM noise in the 40-80 Hz region. This seems to be because 9 MHz RIN coupling to DARM is lower (seen above). Moving around the OMC PZT voltages does not affect the 9 MHz RIN to DARM coupling. Also the OMC LSC dither amplitude was increased from 375 to 6000, 3000, and 1500 cts. This also did not affect the 9 MHz RIN to DARM coupling. We have left it increased to 1500 cts due to the reduction seen in the 180 Hz shoulders. Simple DARM Noisebudget After taking the DCPD cross spectral densities I began wondering what limited the CSD. I retook the intensity noise projection into DARM based on the extra 12 dB of suppression added to ISS second loop, and plotted the intensity projection, frequency projection, DARM correlated noise with 850 averages, and DARM ASD with 2 dB squeezing. (The correlated noise spectrum was spoiled by squeezing, so I have used the non-squeeze CSD of DCPD A and B) There is a distinct shift in the correlated spectrum at around 200 Hz. Above 200 Hz, I think we see only the remainder shot noise of the imperfectly balanced DCPDs. With 850 averages and perfect balance, we should be able to find resolve coherent power at around 1/850 ~ 0.001, but we are able to achieve an DARM ASD ratio of around 1/2 for DARM CSD/DARM ASD, which is (1/2)^2 ~ 0.25 coherence. We ought to be able to balance our PDs a bit better. At 5000 Hz we see some approximately equal combo of intensity and frequency noise. Both are low enough to not limit us even with 2 dB of squeezing. Below 200 Hz, correlated noise dominates. Today this correlated noise between 40 and 80 Hz was slightly better due to this 9 MHz RIN coupling improvement that we don't understand. 30 Hz and below is likely mostly ASC noise. Probably the most important mystery is what is going on at 100 Hz. There is some shelf of correlated noise DARM is sitting on that gives way around 200 Hz. We see some improvement in DARM at 100 Hz from squeezing, so at least some of the noise there is shot noise. With squeezing we are revealing more of the true limiting noise here.
