Elenna Capote, Camilla Compton, Sheila Dwyer, Derek Davis
This afternoon we had a repeat of the bad low frequency noise that we have been suspecting was from filter cavity backscatter 86596. We saw that the symptom of elevated noise in the filter cavity error signal was similar to previous incidents plot.
We compared squeezing with and without the filter cavity, and no squeezing, and see that this noise is there when squeezing is injected no matter what the filter cavity state is. plot and plot with mean sqz and anti squeezing.
We repeated the fringe wrapping measurements, we saw a higher scattered amplitude when moving ZM5 than last week. (shelf is higher by 10dB). the ZM2 shelf is about the same. plot
We also did some 30 Hz excitations in ZM5 + ZM2, we can see a bilinear coupling of these but the background didn't change during this excitation. plot
Derek and Elenna looked at the glitches in DARM that showed up at the time of the noise. Derek ran some hveto runs for times with frequency dependent squeezing and frequecy independent squeezing, and saw that filter cavity length signals are a good witness when the filter cavity is locked, when the filter cavity is not locked the giltches stay but aren't witnessed by the FC error signal.
Camilla found that she could reproduceably make the noise go away by moving the ZM4 +5 PSAMs small amounts. She moved the PSAMs and adjusted the alignment to get a good level of high frequency squeezing back. She also tried to do this with alignment only.
We also took SQZ_OPO_LR GRD to LOCKED_CLF no ISS to check that the pump AOM wasn't injecting any noise.
Ended up leaving ZM4 PSAMs at 6.2V on the strain gauge, old nominal was 6.0V sdf attached. This is only a 9V change on the 0-200V PSAMs, from 78V to 87V. Which doesn't seem big enough to cause such an effect.
- Plot of repeatably changing ZM5 PSAMS along with ZM5 PIT alignment to make noise appear and go
- Plot of repeatably changing ZM4 PSAMS (no ZM4 PIT alignment needed) to make noise appear and go.
- Interestingly 6.0V would be noisy, 6.5V would be noisy, but 6.2V was not! This is confusing.
- Plot of changing ZM4 PIT and YAW alignments slowly ~50urad each way, no difference in noise.
- I also checked ZM5 more quickly and noticed no obvious change
- For ZM4 I tried taking the PIT alignment back to the same osems the good 6.2V PSAMS setting has the osems at, the noise did not change.
Sheila posted backscatter measurements in 86836, and opened an FRS ticket for this issue: FRS # 35457.
B:BS1 is a 99/1 BS (see D2000021 spreadsheet). The PD that the beam transmitting B:BS1 goes to H1:IOO-OFI_PD_A_DC_POWER, this doesn't see any increased noise at the noisy time, although it's only a 16Hz channel: time series and dtt attached.
I could increase the power on B:PD1 by ~25% by moving ZM4 and ZM5 in yaw before we lost RF3 and SQZ went down, showing we are nominally clipping this PD, plot attached. We can repeat Sheila's backscatter measurements with a different amount of light on this PD to see if it's the culprit.
