Reports until 12:22, Tuesday 04 January 2022
H1 SUS (CDS, FRS, SUS)
jeffrey.kissel@LIGO.ORG - posted 12:22, Tuesday 04 January 2022 - last comment - 13:22, Tuesday 04 January 2022(61205)
Replacing SUSH56's ADC0 Card fixes SUS SR3 M1 RT OSEM's Dead/Drifting Channel Issue 
D. Barker, F. Clara, C. Gray, J. Kissel, R. Kumar
FRS Ticket 21820
WP 10144

After Fil narrowed down the dead/drifing channel issue with SUS SR3's M1 (TOP) RT OSEM to the IO chassis (see investigation in LHO aLOG 61196), we tried several things this morning:
   (After Corey and I brought all SUS on sush56 to their SAFE state -- see Austin's D2100766 for the list -- and Dave successfully brought the sush56 computer out of the dolphin network)

    - Dave stopped and shutdown all USER and IOP model processes on the front-end computer, Fil manually power cycled the IO chassis in the CER, then manually turned on the sush56 front-end in the MER, then Fil brought all IOP and USER models back up. In this, we hoped to trigger another autocal sequence. This *did* trigger an autocal, but ADC0 still failed the autocal.

    - Resigned, Fil replaced the ADC card (and ADC Card to ADC Adapter CARD SCSI cable) with a spare -- repeating the shutdown and restart procedure above. This new card *passed* its autocal procedure.

Replacing the ADC card fixed the issue, and all SR3's channels have returned to their pre-Dec 14th non-drifting, much less noisy, normal signals. Upon a cursory glance, all other channels on ADC0 also look good (SR3, SRM, OMC, and OFI M1 stages).

Good work team! 

This aLOG should serve to close FRS Ticket 21820.
Comments related to this report
jeffrey.kissel@LIGO.ORG - 12:23, Tuesday 04 January 2022 (61206)CDS, FRS
Link 

As a side note, we may have identified a bug in the RCG code that handles the DAC DuoTone readback. 

This is the loopback that may be turned on when one hits H1:FEC-43_DACDT_ENABLE switch, which then turns the output of 1st channel of the 1st DAC (DAC_0_0) into a mimic of the timing system's DuoTone signal, and pipes it into the 2nd to last channel of the 1st ADC card (ADC_0_30). Dave recalls that when this is done, typical values read via MEDM screen are oscillating around 0 with peaks in the low thousands (~3500 ct), and *always* *less* than the ADC DuoTone signal that's *always* piped in on the last channel of the ADC (ADC_0_31). 

However, when we hit the button on sush56, we got oscillations that were in the low *10s* of thousands, which Dave argues is impossible. But, then he mused that it might be a calibration issue within the channel -- some 16-bit vs 18-bit vs 20-bit DAC gain difference that is either unaccounted for within the RCG's conversation of this channel, or he's just not used to looking at 18 bit DAC DuoTone readbacks... or something. 

Perhaps worth a small side investigation.
david.barker@LIGO.ORG - 13:22, Tuesday 04 January 2022 (61210)
Link

ADC serial numbers

original ADC (fail autocal, bad ch4) [Removed] 110128-06
first replacement (fail autocal, all chans suspect) [Removed] 110125-06
second replacement (passed autocal) [Installed] 181013-17