In following up the beam tube glitch injections, we by chance noticed that the amplitude spectrum of the NSUM of the ASC photodiodes at end X have changed significantly since ER7. In particular, through July and August there is now a set of peaks at 9 and 10.2Hz and broadband noise in ASC-X_TR that was not there in ER7. We don't see any significant changes in the corresponding Y_TR channels.
The UTC times for the following four plots are:
2015-08-08 14:00:00 (Blue) ("now")
2015-08-02 10:00:00 (Red)
2015-08-01 11:00:00 (Green)
2015-07-31 07:30:00 (Yellow)
2015-06-11 14:00:00 (Purple) ("ER7")
Fig 1: ASD of X_TR_A_NSUM. (excess noise above 5 Hz)
Fig 2: ASD of X_TR_B_NSUM. (excess noise above 5 Hz)
Fig 3: ASD of Y_TR_A_NSUM. (no major changes)
Fig 3: ASD of Y_TR_A_NSUM. (no major changes)
We note that this additional noise should be above the feedback frequency and indeed we see no increased coherence in DARM now, w/r/t ER7, except at 1.7Hz.
Fig 4: TR Coherence w/ DARM ER7 time.
Fig 5: TR Coherence w/ DARM now time.
We thought the most likely explanation would be poor alignment on the TR X ASC diode, coupling beam pointing noise into “fake” intensity noise. However, we don’t see anything obviously wrong with the DC values from the segments now.
Fig 6: Timeseries of TR segments ER7 time. (More light on 1 and 4)
Fig 7: Timeseries TR segments now time. (More light on 3 and 4)
We didn’t see this as a DQ problem anywhere, but wanted to report it in case it indicates some problem we haven’t thought of.