The search for dewar glitches during S4 was done for both sites using the following search criteria:
--Coincidence between SEISZ and either SEISX or SEISY at each location depending on dewar orientation (ie. To detect glitches from a dewar located in the x-direction relative to sensors, use the SEISX and SEISZ channels.)
--Locations/channel combinations:
| LHO | LLO |
|---|---|
| LVEA-X (H0:PEM-LVEA_SEISX & H0:PEM-LVEA_SEISZ) | LVEA-X (L0:PEM-LVEA_SEISX & L0:PEM-LVEA_SEISZ) |
| LVEA-Y (H0:PEM-LVEA_SEISY & H0:PEM-LVEA_SEISZ) | LVEA-Y (L0:PEM-LVEA_SEISY & L0:PEM-LVEA_SEISZ) |
| EX (H0:PEM-EX_SEISX & H0:PEM-EX_SEISZ) | EX (L0:PEM-EX_SEISX & L0:PEM-EX_SEISZ) |
| EY (H0:PEM-EY_SEISY & H0:PEM-EY_SEISZ) | EY (L0:PEM-EY_SEISY & L0:PEM-EY_SEISZ) |
| MX (H0:PEM-MX_SEISX & H0:PEM-MX_SEISZ) | |
| MY (H0:PEM-MY_SEISY & H0:PEM-MY_SEISZ) |
--Seismometer channels filtered at:
5-25 Hz Bandpass
12-20 Hz Bandpass
(Dewar glitches expected to lie at ~16 Hz). An event is considered a glitch if it exceeds threshold for a specified duration of time after filter is applied on both channels.
--Relative Thresholds:
all SEISX, SEISY channels: 10 sigma
all SEISZ channels: 20 sigma
--Duration Constraint:
above threshold for 0.5-5.0 seconds
A visual inspection of the timeseries for each detected glitch allowed several of them to be eliminated as dewar glitch candidates. If the glitch is clearly not centered near 16 Hz or its rise time is greater than ~1 cycle, then it is not considered a dewar glitch. The table below shows the total number of dewar glitch candidates for each site/location. It also contains links to unfiltered time series and spectral density plots for the largest dewar glitches at each location.
The figures for LLO LVEA glitches include timeseries for the AS_Q channel as well as several accelerometers in the LVEA. Differences in the arrival times to each accelerometer give the relative direction to the source. The strongest LVEA glitch caused lock-loss and was detected in both X and Y seismometer channels. However, it was eliminated as a dewar glitch because it did not come from the direction of either dewar (the timeseries shows the signal reaching BSC2_ACCZ before BSC1_ACCZ). Of the remaining 5 LVEA-X glitches, four come from the direction of the X dewar and one (the one at 793921855) is undetermined. Channels PSL1_ACCZ, ISCT1_ACCZ, and ISCT4_ACCZ were used for source location in the X direction. To see filtered timeseries of all detected LLO glitches (including non-dewar) click here.
| LHO (# of dewar glitches) | LLO (# of dewar glitches) | |
|---|---|---|
| LVEA-X | 0 | 5 figures |
| LVEA-Y | 0 | 0 |
| EX | 0 | 1 figures |
| EY | 16 figures | 1 figures |
| MX | 253 figures | |
| MY | 1 figures |
Obvious candidates for insulation are the LHO Mid-X dewars. Because there is only one set of accelerometers at Mid-X, further experiments will be needed to determine which of the Mid-X dewars are glitching.
List of s4 dewar glitch GPS times ( LHO, LLO)--Refine glitch criteria within dewarMon to reduce the number of non-dewar glitches detected.
--Change to absolute thresholds to allow glitch detection during noisy periods (also gives more consistent measure of glitch strength).
--Incorporate accelerometer study (for source location) into dewarMon.
--Prepare monitor for on-line use