The inspiral range had trended down significantly in the last ~2h of the lock.

- Looking back at the data the spectrum showed a 1/f^2 noise floor between 15 and 150Hz. (plot 1)
- This floor was clearly generated by frequent glitches. Plot 2 shows a filtered time series for the bad time (blue, 2015/08/16 15:05 UTC) and a good time (red, 2015/08/16 13:35 UTC)
  Filter: zpk([0;0;],[1000;1000],1,"n")ellip("BandPass",6,1,80,30,200)

- Sheila noticed that both the M2 and M3 coils of the SRM were hovering around the ominous 65536cts mark, so there is the suspicion that these are SRM glitches. However I couldn't line them up convincingly... Either way, Sheila did the top stage bleed-off today, so we will see...