david.barker@LIGO.ORG - posted 15:00, Monday 08 April 2019 - last comment - 16:24, Monday 08 April 2019(48317)
h1edc DAQ status inadvertently set to 0x2000 (config file CRC mismatch) for 4 minutes
While preparing a new H1EPICS_DAQ.ini file for tomorrow's maintenance (to add the missing h1edc daq check sum counter and status) the h1edc went into a 0x2000 DAQ STATUS. The status was in this state between 13:52 (when a new H1EDC.ini file was generated) and 13:56 PDT (when original file was reverted). I cannot get these times with more accuracy because its this STATUS channel I was adding to the DAQ.
Trending several slow channels which follow the changed channels in H1EDC.ini show no VAL problems. Greg is checking the slow channels' DAQ status in the frames during this time period.
Because of this issue I have turned off the auto-generation of the GRD ini and EDC ini files on h1fescript0. I've opened FRS12685.
The CFC flag for h1edc is latched in its Orange state due to these changes. This will be greened up at the next opportunity.
Unfortunately H1 was in observe mode during this time. Greg has confirmed that all channels belonging the h1edc (dcuid=52) were marked as 0x2000 (configuration change) in the frame. I have confirmed that all the data is good through this period, it was just marked with a non-zero status.
To summarize for DetChar:
All external EDCU channels (H1EDC.ini) were incorrectly marked with a non-zero status, data is actually good and should be used.
First frame: H-H1_R-1238791936-64.gwf [Apr 08 2019 20:51:58 UTC]
Last frame: H-H1_R-1238792192-64.gwf [Apr 08 2019 20:56:14 UTC]