Throughout ER8 to present we have seen many examples of the 45 MHz EOM driver acting up. I did a study of this noise between 11th Sept - 23rd Sept using a BLRMS of H1:LSC-MOD_RF45_AM_CTRL_OUT_DQ and from this study, found a threshold on the BLRMS on which to create a data quality flag (details of exactly what I did can be found on this page). This is the flag the analyses use to completely remove this bad data from their searches (CAT1 veto)

From 12th Sept 00:00 UTC - 26th Sept 03:00 UTC the flag marked 23789 seconds (6.61 hours) of LHO data (2.93% of H1:DMT-ANALYSIS_READY). This corresponds to 17611 seconds (4.89 hours) of coincident data (3.14% of {H1,L1}:DMT_ANALYSIS_READY). 

From 26th Sept 03:00 UTC - 28th Sept 00:00 UTC the flag marked 25992 seconds (7.22 hours) of LHO data (17.5% of H1:DMT-ANALYSIS_READY). This corresponds to 10970 seconds (3.05 hours) of coincident data (12.6 % of  {H1,L1}:DMT_ANALYSIS_READY)

The EOM driver seems to have been acting up more in the last few days - the flag has marked more time as bad in the last few days as 2 weeks of data taking!