Jonathan, Jim, Dave:

Details can be found in the maintenance summary below. The LHO DAQ was writing bad data to science, commissioning, second and minute trend files from 12:39 to 15:21 PDT today. Both h1fw0 and h1fw2 were affected. The data corruption is manifested as 16Hz noise on every channel (looks like channel bleed). Only frame writers are affected, so the real-time data from the NDS did not show this corruption.

The problem was quickly resolved, there is a naming restriction for INI files, they cannot end in the 7 character string "GDS.ini".  The new slow data ini file was  named H1EDCU_GDS.ini. This file was removed and the problem went away.

the attached plot shows the noise as seen in a mid station seismometer channel. One second a full data is shown