PEM channels for the ten loudest inspiral candidates were examined using all the channels available in the reduced data set at LLO in most cases.   As colleagues complete corresponding work on other channel sets, we will collate the work, and see what channels may be useful, alone or in combination, for setting vetoes in the S2 data.  The results for the PEM channels, and sometimes a few others besides, are described below.
 

No. 1

First, the unfiltered signal for AS_Q  at gps 730885394.5.

Signals were found in the accelerometer channels at the X end station, BSC4, and the PSL.
 

At the X end station, there was broadband noise to 200 Hz for about 3 seconds, ending at the time of the AS_Q glitch - probably a coincidence.
Time series   filtered through a 20-115 Hz passband.
Spectrogram with the  vertical axis  running up to 1024 Hz.
The signal on the microphone was a steady periodic signal, with fundamental frequency just below 1 Hz.   A similar pattern was see on the BSC5 (Y end) microphone.

The  PSLaccelerometer did not show anything above background at the time of the AS_Q glitch, but did show two broadband glitches to ~900 Hz  occurring 6 sec and 12 sec after the AS_Q glitch.  These were rather surprising in a science segment, but not strong enough to show up in AS_Q.  The time series for the accelerometer has been run through a 100-300 Hz passband filter.  A spectrogram of the raw data is shown here.  The PSL microphone showed nothing at all except for power lines.
 

No. 2

AS_Q at gps 730669092.7, filtered using  a 100 Hz high pass filter.
There is a very minor glitch in the PSL accelerometer at the same time , but nothing that stands out above the noise, and a causal glitch in the PSL accelerometer normally precedes that in AS_Q by 0.5-0.75 s.  (See the tracking investigation.)  Time series using 100-300 Hz filter.
 

No. 3

AS_Q at filtered using 100 Hz HP.
Nothing at all showed up in any of the PEM channels examined.  (Some of the least likely candidates were dropped by this time.)

No. 4

AS_Q at gps 729439355.4 , filtered using 100 Hz HP.
Again nothing of consequence showed up in the PEM channels examined.

No. 5

AS_Qat gps 733032115.9 has snr 64, but is a very narrow spike that does not show upwhen the channel is filtered at 200 Hz high pass.  No PEM channels were glitching any more than average in the time vicinity of this signal.   Chi-squared = 86.6 - measuring fit against an inspiral waveform.

No. 6

AS_Q at gps 730669107.6, using a 100 Hz HP filter.  A spectrogram is shown here.  Chi-squared = 43.4.
There are bursty glitches in a number of the accelerometer channels: HAM3_ACCZ, HAM4_ACCX, BSC3_ACCZ, AND BSC4_ACCZ.  Only in the last channel was there a burst around the time of the AS_Q glitch, and the magnitudes of preceding glitches in the frame were comparable.

No. 7

AS_Q at gps 730407900.9 using 100 Hz HP filter.   The event candidate has snr 52, chi-squared 48.6, and occurs in a stretch of bursty noise.  A spectrogram shows signal to frequencies above 600 Hz.

The SEISZ channels at the X end station and the LVEA show a lot of noise, and a number of the accelerometer channels in the LVEA are seriously bursty.  This was noted for BSC1-3 and HAM1-3, both horizontal and vertical channels.   Spectrograms show a strong line ~280 Hz in a number of these channels, e.g. HAM2_ACCZ, which also shows broadband noise around 800Hz, and some, e.g. HAM1_ACCX,  show strong broadband noise between 500 and 900 Hz.  Re-examining the seismic Z channelsusing spectrograms, the bursts are stronger at the X end station, but staying below 20 Hz.  At the LVEA, the signals are weaker, but extending up to 30 Hz.  (At ~1 pm local time, there was very likely to be logging or other human activity near the X end station.)  The Streckheinsen seismometer at the X end station recorded a glitch peaking at ~73040900.5.

Moving to interferometer channels, POB_I and POB_Q show clear glitches, as does REFL_Q, but not REFL_I.
 

No. 8

AS_Q at gps 731605243.6 using a 100 Hz HP filter (snr 50 from templates), and using a 150 Hz HP filter.  The signal has disappeared completely in this case.   Not surprisingly, nothing was found in any of the PEM channels (including the microphones).  Chi-squared = 64.2.

No. 9

AS_Q at gps 733535580.2 using 100 Hz HP filter.  The signal appears to be essentially non-existent above 100 Hz, and this is confirmed by a spectrogram.  The signal should show up at 12.2 seconds in this frame.  The reported snr and chi-wquared values were 47 and 77.6 respectively.

No. 10

AS_Qat gps 731605049.8 using 100 Hz HP filter.   This spike is not visible when using 150 Hz HP filter.   A spectrogram confirms that the signal is negligible.   The snr and chi-squared values are 42 and 96.4 respectively.